Authors: Rishika Bhagwatkar, Shravan Nayak, Reza Bayat, Alexis Roger, Daniel Z Kaplan, Pouya Bashivan, Irina Rish
Abstract: Vision-Language Models (VLMs) have witnessed a surge in both research and real-world applications. However, as they are becoming increasingly prevalent, ensuring their robustness against adversarial attacks is paramount. This work systematically investigates the impact of model design choices on the adversarial robustness of VLMs against image-based attacks. Additionally, we introduce novel, cost-effective approaches to enhance robustness through prompt formatting. By rephrasing questions and suggesting potential adversarial perturbations, we demonstrate substantial improvements in model robustness against strong image-based attacks such as Auto-PGD. Our findings provide important guidelines for developing more robust VLMs, particularly for deployment in safety-critical environments.
Authors: Amol S. Joshi, Ali Dabouei, Jeremy Dawson, Nasser Nasrabadi
Abstract: Quality assessment of fingerprints captured using digital cameras and smartphones, also called fingerphotos, is a challenging problem in biometric recognition systems. As contactless biometric modalities are gaining more attention, their reliability should also be improved. Many factors, such as illumination, image contrast, camera angle, etc., in fingerphoto acquisition introduce various types of distortion that may render the samples useless. Current quality estimation methods developed for fingerprints collected using contact-based sensors are inadequate for fingerphotos. We propose Utility guided Fingerphoto Quality Assessment (UFQA), a self-supervised dual encoder framework to learn meaningful feature representations to assess fingerphoto quality. A quality prediction model is trained to assess fingerphoto quality with additional supervision of quality maps. The quality metric is a predictor of the utility of fingerphotos in matching scenarios. Therefore, we use a holistic approach by including fingerphoto utility and local quality when labeling the training data. Experimental results verify that our approach performs better than the widely used fingerprint quality metric NFIQ2.2 and state-of-the-art image quality assessment algorithms on multiple publicly available fingerphoto datasets.
Authors: Yifei Wang, Weimin Bai, Weijian Luo, Wenzheng Chen, He Sun
Abstract: Diffusion models (DMs) have emerged as powerful generative models for solving inverse problems, offering a good approximation of prior distributions of real-world image data. Typically, diffusion models rely on large-scale clean signals to accurately learn the score functions of ground truth clean image distributions. However, such a requirement for large amounts of clean data is often impractical in real-world applications, especially in fields where data samples are expensive to obtain. To address this limitation, in this work, we introduce \emph{FlowDiff}, a novel joint training paradigm that leverages a conditional normalizing flow model to facilitate the training of diffusion models on corrupted data sources. The conditional normalizing flow try to learn to recover clean images through a novel amortized inference mechanism, and can thus effectively facilitate the diffusion model's training with corrupted data. On the other side, diffusion models provide strong priors which in turn improve the quality of image recovery. The flow model and the diffusion model can therefore promote each other and demonstrate strong empirical performances. Our elaborate experiment shows that FlowDiff can effectively learn clean distributions across a wide range of corrupted data sources, such as noisy and blurry images. It consistently outperforms existing baselines with significant margins under identical conditions. Additionally, we also study the learned diffusion prior, observing its superior performance in downstream computational imaging tasks, including inpainting, denoising, and deblurring.
Authors: Ioannis Maniadis Metaxas, Georgios Tzimiropoulos, Ioannis Patras
Abstract: Self-supervised learning has recently emerged as the preeminent pretraining paradigm across and between modalities, with remarkable results. In the image domain specifically, group (or cluster) discrimination has been one of the most successful methods. However, such frameworks need to guard against heavily imbalanced cluster assignments to prevent collapse to trivial solutions. Existing works typically solve this by reweighing cluster assignments to promote balance, or with offline operations (e.g. regular re-clustering) that prevent collapse. However, the former typically requires large batch sizes, which leads to increased resource requirements, and the latter introduces scalability issues with regard to large datasets. In this work, we propose ExCB, a framework that tackles this problem with a novel cluster balancing method. ExCB estimates the relative size of the clusters across batches and balances them by adjusting cluster assignments, proportionately to their relative size and in an online manner. Thereby, it overcomes previous methods' dependence on large batch sizes and is fully online, and therefore scalable to any dataset. We conduct extensive experiments to evaluate our approach and demonstrate that ExCB: a) achieves state-of-the-art results with significantly reduced resource requirements compared to previous works, b) is fully online, and therefore scalable to large datasets, and c) is stable and effective even with very small batch sizes.
Authors: Pramish Paudel, Anubhav Khanal, Ajad Chhatkuli, Danda Pani Paudel, Jyoti Tandukar
Abstract: Personalized 3D avatars require an animatable representation of digital humans. Doing so instantly from monocular videos offers scalability to broad class of users and wide-scale applications. In this paper, we present a fast, simple, yet effective method for creating animatable 3D digital humans from monocular videos. Our method utilizes the efficiency of Gaussian splatting to model both 3D geometry and appearance. However, we observed that naively optimizing Gaussian splats results in inaccurate geometry, thereby leading to poor animations. This work achieves and illustrates the need of accurate 3D mesh-type modelling of the human body for animatable digitization through Gaussian splats. This is achieved by developing a novel pipeline that benefits from three key aspects: (a) implicit modelling of surface's displacements and the color's spherical harmonics; (b) binding of 3D Gaussians to the respective triangular faces of the body template; (c) a novel technique to render normals followed by their auxiliary supervision. Our exhaustive experiments on three different benchmark datasets demonstrates the state-of-the-art results of our method, in limited time settings. In fact, our method is faster by an order of magnitude (in terms of training time) than its closest competitor. At the same time, we achieve superior rendering and 3D reconstruction performance under the change of poses.
Authors: Ekaterina Zaychenkova, Dmitrii Iarchuk, Sergey Korchagin, Alexey Zaitsev, Egor Ershov
Abstract: In medical diagnostics, accurate uncertainty estimation for neural-based models is essential for complementing second-opinion systems. Despite neural network ensembles' proficiency in this problem, a gap persists between actual uncertainties and predicted estimates. A major difficulty here is the lack of labels on the hardness of examples: a typical dataset includes only ground truth target labels, making the uncertainty estimation problem almost unsupervised. Our novel approach narrows this gap by integrating expert assessments of case complexity into the neural network's learning process, utilizing both definitive target labels and supplementary complexity ratings. We validate our methodology for blood typing, leveraging a new dataset "BloodyWell" unique in augmenting labeled reaction images with complexity scores from six medical specialists. Experiments demonstrate enhancement of our approach in uncertainty prediction, achieving a 2.5-fold improvement with expert labels and a 35% increase in performance with estimates of neural-based expert consensus.
Authors: Chenwei Wu, David Restrepo, Zitao Shuai, Zhongming Liu, Liyue Shen
Abstract: In-context learning (ICL) with Large Vision Models (LVMs) presents a promising avenue in medical image segmentation by reducing the reliance on extensive labeling. However, the ICL performance of LVMs highly depends on the choices of visual prompts and suffers from domain shifts. While existing works leveraging LVMs for medical tasks have focused mainly on model-centric approaches like fine-tuning, we study an orthogonal data-centric perspective on how to select good visual prompts to facilitate generalization to medical domain. In this work, we propose a label-efficient in-context medical segmentation method by introducing a novel Meta-driven Visual Prompt Selection mechanism (MVPS), where a prompt retriever obtained from a meta-learning framework actively selects the optimal images as prompts to promote model performance and generalizability. Evaluated on 8 datasets and 4 tasks across 3 medical imaging modalities, our proposed approach demonstrates consistent gains over existing methods under different scenarios, improving both computational and label efficiency. Finally, we show that MVPS is a flexible, finetuning-free module that could be easily plugged into different backbones and combined with other model-centric approaches.
Authors: Vijul Shah, Ko Watanabe, Brian B. Moser, Andreas Dengel
Abstract: In this work, we introduce EyeDentify, a dataset specifically designed for pupil diameter estimation based on webcam images. EyeDentify addresses the lack of available datasets for pupil diameter estimation, a crucial domain for understanding physiological and psychological states traditionally dominated by highly specialized sensor systems such as Tobii. Unlike these advanced sensor systems and associated costs, webcam images are more commonly found in practice. Yet, deep learning models that can estimate pupil diameters using standard webcam data are scarce. By providing a dataset of cropped eye images alongside corresponding pupil diameter information, EyeDentify enables the development and refinement of models designed specifically for less-equipped environments, democratizing pupil diameter estimation by making it more accessible and broadly applicable, which in turn contributes to multiple domains of understanding human activity and supporting healthcare. Our dataset is available at https://vijulshah.github.io/eyedentify/.
Authors: Philipp Allgeuer, Kyra Ahrens, Stefan Wermter
Abstract: We introduce NOVIC, an innovative uNconstrained Open Vocabulary Image Classifier that uses an autoregressive transformer to generatively output classification labels as language. Leveraging the extensive knowledge of CLIP models, NOVIC harnesses the embedding space to enable zero-shot transfer from pure text to images. Traditional CLIP models, despite their ability for open vocabulary classification, require an exhaustive prompt of potential class labels, restricting their application to images of known content or context. To address this, we propose an "object decoder" model that is trained on a large-scale 92M-target dataset of templated object noun sets and LLM-generated captions to always output the object noun in question. This effectively inverts the CLIP text encoder and allows textual object labels to be generated directly from image-derived embedding vectors, without requiring any a priori knowledge of the potential content of an image. The trained decoders are tested on a mix of manually and web-curated datasets, as well as standard image classification benchmarks, and achieve fine-grained prompt-free prediction scores of up to 87.5%, a strong result considering the model must work for any conceivable image and without any contextual clues.
Authors: Zijian Zhou, Zheng Zhu, Holger Caesar, Miaojing Shi
Abstract: Panoptic Scene Graph Generation (PSG) aims to segment objects and recognize their relations, enabling the structured understanding of an image. Previous methods focus on predicting predefined object and relation categories, hence limiting their applications in the open world scenarios. With the rapid development of large multimodal models (LMMs), significant progress has been made in open-set object detection and segmentation, yet open-set relation prediction in PSG remains unexplored. In this paper, we focus on the task of open-set relation prediction integrated with a pretrained open-set panoptic segmentation model to achieve true open-set panoptic scene graph generation (OpenPSG). Our OpenPSG leverages LMMs to achieve open-set relation prediction in an autoregressive manner. We introduce a relation query transformer to efficiently extract visual features of object pairs and estimate the existence of relations between them. The latter can enhance the prediction efficiency by filtering irrelevant pairs. Finally, we design the generation and judgement instructions to perform open-set relation prediction in PSG autoregressively. To our knowledge, we are the first to propose the open-set PSG task. Extensive experiments demonstrate that our method achieves state-of-the-art performance in open-set relation prediction and panoptic scene graph generation. Code is available at \url{https://github.com/franciszzj/OpenPSG}.
Authors: Hoonhee Cho, Sung-Hoon Yoon, Hyeokjun Kweon, Kuk-Jin Yoon
Abstract: Event cameras excel in capturing high-contrast scenes and dynamic objects, offering a significant advantage over traditional frame-based cameras. Despite active research into leveraging event cameras for semantic segmentation, generating pixel-wise dense semantic maps for such challenging scenarios remains labor-intensive. As a remedy, we present EV-WSSS: a novel weakly supervised approach for event-based semantic segmentation that utilizes sparse point annotations. To fully leverage the temporal characteristics of event data, the proposed framework performs asymmetric dual-student learning between 1) the original forward event data and 2) the longer reversed event data, which contain complementary information from the past and the future, respectively. Besides, to mitigate the challenges posed by sparse supervision, we propose feature-level contrastive learning based on class-wise prototypes, carefully aggregated at both spatial region and sample levels. Additionally, we further excavate the potential of our dual-student learning model by exchanging prototypes between the two learning paths, thereby harnessing their complementary strengths. With extensive experiments on various datasets, including DSEC Night-Point with sparse point annotations newly provided by this paper, the proposed method achieves substantial segmentation results even without relying on pixel-level dense ground truths. The code and dataset are available at https://github.com/Chohoonhee/EV-WSSS.
Authors: Nian Wu, Jiarui Xing, Miaomiao Zhang
Abstract: This paper presents a novel approach, termed {\em Temporal Latent Residual Network (TLRN)}, to predict a sequence of deformation fields in time-series image registration. The challenge of registering time-series images often lies in the occurrence of large motions, especially when images differ significantly from a reference (e.g., the start of a cardiac cycle compared to the peak stretching phase). To achieve accurate and robust registration results, we leverage the nature of motion continuity and exploit the temporal smoothness in consecutive image frames. Our proposed TLRN highlights a temporal residual network with residual blocks carefully designed in latent deformation spaces, which are parameterized by time-sequential initial velocity fields. We treat a sequence of residual blocks over time as a dynamic training system, where each block is designed to learn the residual function between desired deformation features and current input accumulated from previous time frames. We validate the effectivenss of TLRN on both synthetic data and real-world cine cardiac magnetic resonance (CMR) image videos. Our experimental results shows that TLRN is able to achieve substantially improved registration accuracy compared to the state-of-the-art. Our code is publicly available at https://github.com/nellie689/TLRN.
Authors: Adam Korycki, Colleen Josephson, Steve McGuire
Abstract: As Neural Radiance Field (NeRF) implementations become faster, more efficient and accurate, their applicability to real world mapping tasks becomes more accessible. Traditionally, 3D mapping, or scene reconstruction, has relied on expensive LiDAR sensing. Photogrammetry can perform image-based 3D reconstruction but is computationally expensive and requires extremely dense image representation to recover complex geometry and photorealism. NeRFs perform 3D scene reconstruction by training a neural network on sparse image and pose data, achieving superior results to photogrammetry with less input data. This paper presents an evaluation of two NeRF scene reconstructions for the purpose of estimating the diameter of a vertical PVC cylinder. One of these are trained on commodity iPhone data and the other is trained on robot-sourced imagery and poses. This neural-geometry is compared to state-of-the-art lidar-inertial SLAM in terms of scene noise and metric-accuracy.
Authors: Mang Ning, Albert Ali Salah, Itir Onal Ertugrul
Abstract: Facial Action Unit (AU) detection has gained significant research attention as AUs contain complex expression information. In this paper, we unpack two fundamental factors in AU detection: data and subject identity regularization, respectively. Motivated by recent advances in foundation models, we highlight the importance of data and collect a diverse dataset Face9M, comprising 9 million facial images, from multiple public resources. Pretraining a masked autoencoder on Face9M yields strong performance in AU detection and facial expression tasks. We then show that subject identity in AU datasets provides a shortcut learning for the model and leads to sub-optimal solutions to AU predictions. To tackle this generic issue of AU tasks, we propose Identity Adversarial Training (IAT) and demonstrate that a strong IAT regularization is necessary to learn identity-invariant features. Furthermore, we elucidate the design space of IAT and empirically show that IAT circumvents the identity shortcut learning and results in a better solution. Our proposed methods, Facial Masked Autoencoder (FMAE) and IAT, are simple, generic and effective. Remarkably, the proposed FMAE-IAT approach achieves new state-of-the-art F1 scores on BP4D (67.1\%), BP4D+ (66.8\%), and DISFA (70.1\%) databases, significantly outperforming previous work. We release the code and model at https://github.com/forever208/FMAE-IAT, the first open-sourced facial model pretrained on 9 million diverse images.
Authors: Rong Wang, Wei Mao, Changsheng Lu, Hongdong Li
Abstract: Animating stylized characters to match a reference motion sequence is a highly demanded task in film and gaming industries. Existing methods mostly focus on rigid deformations of characters' body, neglecting local deformations on the apparel driven by physical dynamics. They deform apparel the same way as the body, leading to results with limited details and unrealistic artifacts, e.g. body-apparel penetration. In contrast, we present a novel method aiming for high-quality motion transfer with realistic apparel animation. As existing datasets lack annotations necessary for generating realistic apparel animations, we build a new dataset named MMDMC, which combines stylized characters from the MikuMikuDance community with real-world Motion Capture data. We then propose a data-driven pipeline that learns to disentangle body and apparel deformations via two neural deformation modules. For body parts, we propose a geodesic attention block to effectively incorporate semantic priors into skeletal body deformation to tackle complex body shapes for stylized characters. Since apparel motion can significantly deviate from respective body joints, we propose to model apparel deformation in a non-linear vertex displacement field conditioned on its historic states. Extensive experiments show that our method produces results with superior quality for various types of apparel. Our dataset is released in https://github.com/rongakowang/MMDMC.
Authors: Yihao Luo, Yikai Wang, Zhengrui Xiang, Yuliang Xiu, Guang Yang, ChoonHwai Yap
Abstract: In this paper, we propose the differentiable voxelization of 3D meshes via the winding number and solid angles. The proposed approach achieves fast, flexible, and accurate voxelization of 3D meshes, admitting the computation of gradients with respect to the input mesh and GPU acceleration. We further demonstrate the application of the proposed voxelization in mesh morphing, where the voxelized mesh is deformed by a neural network. The proposed method is evaluated on the ShapeNet dataset and achieves state-of-the-art performance in terms of both accuracy and efficiency.
Authors: Abdollah Zakeri, Mulham Fawakherji, Jiming Kang, Bikram Koirala, Venkatesh Balan, Weihang Zhu, Driss Benhaddou, Fatima A. Merchant
Abstract: Automating agricultural processes holds significant promise for enhancing efficiency and sustainability in various farming practices. This paper contributes to the automation of agricultural processes by providing a dedicated mushroom detection dataset related to automated harvesting, growth monitoring, and quality control of the button mushroom produced using Agaricus Bisporus fungus. With over 18,000 mushroom instances in 423 RGB-D image pairs taken with an Intel RealSense D405 camera, it fills the gap in mushroom-specific datasets and serves as a benchmark for detection and instance segmentation algorithms in smart mushroom agriculture. The dataset, featuring realistic growth environment scenarios with comprehensive annotations, is assessed using advanced detection and instance segmentation algorithms. The paper details the dataset's characteristics, evaluates algorithmic performance, and for broader applicability, we have made all resources publicly available including images, codes, and trained models via our GitHub repository https://github.com/abdollahzakeri/m18k
Authors: Guanghua He, Wangang Cheng, Hancan Zhu, Gaohang Yu
Abstract: The hippocampus is a crucial brain structure associated with various psychiatric disorders, and its automatic and precise segmentation is essential for studying these diseases. In recent years, deep learning-based methods have made significant progress in hippocampus segmentation. However, training deep neural network models requires substantial computational resources and time, as well as a large amount of labeled training data, which is often difficult to obtain in medical image segmentation. To address this issue, we propose a new parameter-efficient fine-tuning method called LoRA-PT. This method transfers the pre-trained UNETR model on the BraTS2021 dataset to the hippocampus segmentation task. Specifically, the LoRA-PT method categorizes the parameter matrix of the transformer structure into three sizes, forming three 3D tensors. Through tensor singular value decomposition, these tensors are decomposed to generate low-rank tensors with the principal singular values and singular vectors, while the remaining singular values and vectors form the residual tensor. Similar to the LoRA method, during parameter fine-tuning, we only update the low-rank tensors, i.e. the principal tensor singular values and vectors, while keeping the residual tensor unchanged. We validated the proposed method on three public hippocampus datasets. Experimental results show that LoRA-PT outperforms existing parameter-efficient transfer learning methods in segmentation accuracy while significantly reducing the number of parameter updates.
Authors: Liu He, Daniel Aliaga
Abstract: The generation of large-scale urban layouts has garnered substantial interest across various disciplines. Prior methods have utilized procedural generation requiring manual rule coding or deep learning needing abundant data. However, prior approaches have not considered the context-sensitive nature of urban layout generation. Our approach addresses this gap by leveraging a canonical graph representation for the entire city, which facilitates scalability and captures the multi-layer semantics inherent in urban layouts. We introduce a novel graph-based masked autoencoder (GMAE) for city-scale urban layout generation. The method encodes attributed buildings, city blocks, communities and cities into a unified graph structure, enabling self-supervised masked training for graph autoencoder. Additionally, we employ scheduled iterative sampling for 2.5D layout generation, prioritizing the generation of important city blocks and buildings. Our approach achieves good realism, semantic consistency, and correctness across the heterogeneous urban styles in 330 US cities. Codes and datasets are released at https://github.com/Arking1995/COHO.
Authors: Yuxuan Lei, Dingkang Yang, Zhaoyu Chen, Jiawei Chen, Peng Zhai, Lihua Zhang
Abstract: Context-aware emotion recognition (CAER) is a complex and significant task that requires perceiving emotions from various contextual cues. Previous approaches primarily focus on designing sophisticated architectures to extract emotional cues from images. However, their knowledge is confined to specific training datasets and may reflect the subjective emotional biases of the annotators. Furthermore, acquiring large amounts of labeled data is often challenging in real-world applications. In this paper, we systematically explore the potential of leveraging Large Vision-Language Models (LVLMs) to empower the CAER task from three paradigms: 1) We fine-tune LVLMs on two CAER datasets, which is the most common way to transfer large models to downstream tasks. 2) We design zero-shot and few-shot patterns to evaluate the performance of LVLMs in scenarios with limited data or even completely unseen. In this case, a training-free framework is proposed to fully exploit the In-Context Learning (ICL) capabilities of LVLMs. Specifically, we develop an image similarity-based ranking algorithm to retrieve examples; subsequently, the instructions, retrieved examples, and the test example are combined to feed LVLMs to obtain the corresponding sentiment judgment. 3) To leverage the rich knowledge base of LVLMs, we incorporate Chain-of-Thought (CoT) into our framework to enhance the model's reasoning ability and provide interpretable results. Extensive experiments and analyses demonstrate that LVLMs achieve competitive performance in the CAER task across different paradigms. Notably, the superior performance in few-shot settings indicates the feasibility of LVLMs for accomplishing specific tasks without extensive training.
Authors: Pierre-David Letourneau, Manish Kumar Singh, Hsin-Pai Cheng, Shizhong Han, Yunxiao Shi, Dalton Jones, Matthew Harper Langston, Hong Cai, Fatih Porikli
Abstract: We present Polynomial Attention Drop-in Replacement (PADRe), a novel and unifying framework designed to replace the conventional self-attention mechanism in transformer models. Notably, several recent alternative attention mechanisms, including Hyena, Mamba, SimA, Conv2Former, and Castling-ViT, can be viewed as specific instances of our PADRe framework. PADRe leverages polynomial functions and draws upon established results from approximation theory, enhancing computational efficiency without compromising accuracy. PADRe's key components include multiplicative nonlinearities, which we implement using straightforward, hardware-friendly operations such as Hadamard products, incurring only linear computational and memory costs. PADRe further avoids the need for using complex functions such as Softmax, yet it maintains comparable or superior accuracy compared to traditional self-attention. We assess the effectiveness of PADRe as a drop-in replacement for self-attention across diverse computer vision tasks. These tasks include image classification, image-based 2D object detection, and 3D point cloud object detection. Empirical results demonstrate that PADRe runs significantly faster than the conventional self-attention (11x ~ 43x faster on server GPU and mobile NPU) while maintaining similar accuracy when substituting self-attention in the transformer models.
Authors: Liuyue Xie, Joel Julin, Koichiro Niinuma, Laszlo A. Jeni
Abstract: Reconstructing dynamic 3D scenes from 2D images and generating diverse views over time presents a significant challenge due to the inherent complexity and temporal dynamics involved. While recent advancements in neural implicit models and dynamic Gaussian Splatting have shown promise, limitations persist, particularly in accurately capturing the underlying geometry of highly dynamic scenes. Some approaches address this by incorporating strong semantic and geometric priors through diffusion models. However, we explore a different avenue by investigating the potential of regularizing the native warp field within the dynamic Gaussian Splatting framework. Our method is grounded on the key intuition that an accurate warp field should produce continuous space-time motions. While enforcing the motion constraints on warp fields is non-trivial, we show that we can exploit knowledge innate to the forward warp field network to derive an analytical velocity field, then time integrate for scene flows to effectively constrain both the 2D motion and 3D positions of the Gaussians. This derived Lucas-Kanade style analytical regularization enables our method to achieve superior performance in reconstructing highly dynamic scenes, even under minimal camera movement, extending the boundaries of what existing dynamic Gaussian Splatting frameworks can achieve.
Authors: Wang Zeng, Sheng Jin, Lumin Xu, Wentao Liu, Chen Qian, Wanli Ouyang, Ping Luo, Xiaogang Wang
Abstract: Transformers are widely used in computer vision areas and have achieved remarkable success. Most state-of-the-art approaches split images into regular grids and represent each grid region with a vision token. However, fixed token distribution disregards the semantic meaning of different image regions, resulting in sub-optimal performance. To address this issue, we propose the Token Clustering Transformer (TCFormer), which generates dynamic vision tokens based on semantic meaning. Our dynamic tokens possess two crucial characteristics: (1) Representing image regions with similar semantic meanings using the same vision token, even if those regions are not adjacent, and (2) concentrating on regions with valuable details and represent them using fine tokens. Through extensive experimentation across various applications, including image classification, human pose estimation, semantic segmentation, and object detection, we demonstrate the effectiveness of our TCFormer. The code and models for this work are available at https://github.com/zengwang430521/TCFormer.
Authors: Cilin Yan, Haochen Wang, Shilin Yan, Xiaolong Jiang, Yao Hu, Guoliang Kang, Weidi Xie, Efstratios Gavves
Abstract: Existing Video Object Segmentation (VOS) relies on explicit user instructions, such as categories, masks, or short phrases, restricting their ability to perform complex video segmentation requiring reasoning with world knowledge. In this paper, we introduce a new task, Reasoning Video Object Segmentation (ReasonVOS). This task aims to generate a sequence of segmentation masks in response to implicit text queries that require complex reasoning abilities based on world knowledge and video contexts, which is crucial for structured environment understanding and object-centric interactions, pivotal in the development of embodied AI. To tackle ReasonVOS, we introduce VISA (Video-based large language Instructed Segmentation Assistant), to leverage the world knowledge reasoning capabilities of multi-modal LLMs while possessing the ability to segment and track objects in videos with a mask decoder. Moreover, we establish a comprehensive benchmark consisting of 35,074 instruction-mask sequence pairs from 1,042 diverse videos, which incorporates complex world knowledge reasoning into segmentation tasks for instruction-tuning and evaluation purposes of ReasonVOS models. Experiments conducted on 8 datasets demonstrate the effectiveness of VISA in tackling complex reasoning segmentation and vanilla referring segmentation in both video and image domains. The code and dataset are available at https://github.com/cilinyan/VISA.
Authors: Penghui Du, Yu Wang, Yifan Sun, Luting Wang, Yue Liao, Gang Zhang, Errui Ding, Yan Wang, Jingdong Wang, Si Liu
Abstract: Existing methods enhance open-vocabulary object detection by leveraging the robust open-vocabulary recognition capabilities of Vision-Language Models (VLMs), such as CLIP.However, two main challenges emerge:(1) A deficiency in concept representation, where the category names in CLIP's text space lack textual and visual knowledge.(2) An overfitting tendency towards base categories, with the open vocabulary knowledge biased towards base categories during the transfer from VLMs to detectors.To address these challenges, we propose the Language Model Instruction (LaMI) strategy, which leverages the relationships between visual concepts and applies them within a simple yet effective DETR-like detector, termed LaMI-DETR.LaMI utilizes GPT to construct visual concepts and employs T5 to investigate visual similarities across categories.These inter-category relationships refine concept representation and avoid overfitting to base categories.Comprehensive experiments validate our approach's superior performance over existing methods in the same rigorous setting without reliance on external training resources.LaMI-DETR achieves a rare box AP of 43.4 on OV-LVIS, surpassing the previous best by 7.8 rare box AP.
Authors: Yunhui Han, Kun Yu, Zhiwei Li
Abstract: Lane topology, which is usually modeled by a centerline graph, is essential for high-level autonomous driving. For a high-quality graph, both topology connectivity and spatial continuity of centerline segments are critical. However, most of existing approaches pay more attention to connectivity while neglect the continuity. Such kind of centerline graph usually cause problem to planning of autonomous driving. To overcome this problem, we present an end-to-end network, CGNet, with three key modules: 1)Junction Aware Query Enhancement module, which provides positional prior to accurately predict junction points; 2)B\'ezier Space Connection module, which enforces continuity constraints on any two topologically connected segments in a B\'ezier space; 3) Iterative Topology Refinement module, which is a graph-based network with memory to iteratively refine the predicted topological connectivity. CGNet achieves state-of-the-art performance on both nuScenes and Argoverse2 datasets.
Authors: Jingqian Wu, Shuo Zhu, Chutian Wang, Edmund Y. Lam
Abstract: Computational neuromorphic imaging (CNI) with event cameras offers advantages such as minimal motion blur and enhanced dynamic range, compared to conventional frame-based methods. Existing event-based radiance field rendering methods are built on neural radiance field, which is computationally heavy and slow in reconstruction speed. Motivated by the two aspects, we introduce Ev-GS, the first CNI-informed scheme to infer 3D Gaussian splatting from a monocular event camera, enabling efficient novel view synthesis. Leveraging 3D Gaussians with pure event-based supervision, Ev-GS overcomes challenges such as the detection of fast-moving objects and insufficient lighting. Experimental results show that Ev-GS outperforms the method that takes frame-based signals as input by rendering realistic views with reduced blurring and improved visual quality. Moreover, it demonstrates competitive reconstruction quality and reduced computing occupancy compared to existing methods, which paves the way to a highly efficient CNI approach for signal processing.
Authors: Xu Zheng, Yuanhuiyi Lyu, Jiazhou Zhou, Lin Wang
Abstract: Fusing an arbitrary number of modalities is vital for achieving robust multi-modal fusion of semantic segmentation yet remains less explored to date. Recent endeavors regard RGB modality as the center and the others as the auxiliary, yielding an asymmetric architecture with two branches. However, the RGB modality may struggle in certain circumstances, e.g., nighttime, while others, e.g., event data, own their merits; thus, it is imperative for the fusion model to discern robust and fragile modalities, and incorporate the most robust and fragile ones to learn a resilient multi-modal framework. To this end, we propose a novel method, named MAGIC, that can be flexibly paired with various backbones, ranging from compact to high-performance models. Our method comprises two key plug-and-play modules. Firstly, we introduce a multi-modal aggregation module to efficiently process features from multi-modal batches and extract complementary scene information. On top, a unified arbitrary-modal selection module is proposed to utilize the aggregated features as the benchmark to rank the multi-modal features based on the similarity scores. This way, our method can eliminate the dependence on RGB modality and better overcome sensor failures while ensuring the segmentation performance. Under the commonly considered multi-modal setting, our method achieves state-of-the-art performance while reducing the model parameters by 60%. Moreover, our method is superior in the novel modality-agnostic setting, where it outperforms prior arts by a large margin of +19.41% mIoU
Authors: Gwangtak Bae, Changwoon Choi, Hyeongjun Heo, Sang Min Kim, Young Min Kim
Abstract: We present an inverse image-formation module that can enhance the robustness of existing visual SLAM pipelines for casually captured scenarios. Casual video captures often suffer from motion blur and varying appearances, which degrade the final quality of coherent 3D visual representation. We propose integrating the physical imaging into the SLAM system, which employs linear HDR radiance maps to collect measurements. Specifically, individual frames aggregate images of multiple poses along the camera trajectory to explain prevalent motion blur in hand-held videos. Additionally, we accommodate per-frame appearance variation by dedicating explicit variables for image formation steps, namely white balance, exposure time, and camera response function. Through joint optimization of additional variables, the SLAM pipeline produces high-quality images with more accurate trajectories. Extensive experiments demonstrate that our approach can be incorporated into recent visual SLAM pipelines using various scene representations, such as neural radiance fields or Gaussian splatting.
Authors: Seo-Bin Hwang, Han-Young Kim, Chae-Yeon Heo, Hie-Yong Jung, Sung-Ju Jung, Yeong-Jun Cho
Abstract: The flatfish is a major farmed species consumed globally in large quantities. However, due to the densely populated farming environment, flatfish are susceptible to injuries and diseases, making early disease detection crucial. Traditionally, diseases were detected through visual inspection, but observing large numbers of fish is challenging. Automated approaches based on deep learning technologies have been widely used, to address this problem, but accurate detection remains difficult due to the diversity of the fish and the lack of the fish disease dataset. In this study, augments fish disease images using generative adversarial networks and image harmonization methods. Next, disease detectors are trained separately for three body parts (head, fins, and body) to address individual diseases properly. In addition, a flatfish disease image dataset called \texttt{FlatIMG} is created and verified on the dataset using the proposed methods. A flash salmon disease dataset is also tested to validate the generalizability of the proposed methods. The results achieved 12\% higher performance than the baseline framework. This study is the first attempt to create a large-scale flatfish disease image dataset and propose an effective disease detection framework. Automatic disease monitoring could be achieved in farming environments based on the proposed methods and dataset.
Authors: Xu Zheng, Yuanhuiyi Lyu, Lin Wang
Abstract: Image modality is not perfect as it often fails in certain conditions, e.g., night and fast motion. This significantly limits the robustness and versatility of existing multi-modal (i.e., Image+X) semantic segmentation methods when confronting modality absence or failure, as often occurred in real-world applications. Inspired by the open-world learning capability of multi-modal vision-language models (MVLMs), we explore a new direction in learning the modality-agnostic representation via knowledge distillation (KD) from MVLMs. Intuitively, we propose Any2Seg, a novel framework that can achieve robust segmentation from any combination of modalities in any visual conditions. Specifically, we first introduce a novel language-guided semantic correlation distillation (LSCD) module to transfer both inter-modal and intra-modal semantic knowledge in the embedding space from MVLMs, e.g., LanguageBind. This enables us to minimize the modality gap and alleviate semantic ambiguity to combine any modalities in any visual conditions. Then, we introduce a modality-agnostic feature fusion (MFF) module that reweights the multi-modal features based on the inter-modal correlation and selects the fine-grained feature. This way, our Any2Seg finally yields an optimal modality-agnostic representation. Extensive experiments on two benchmarks with four modalities demonstrate that Any2Seg achieves the state-of-the-art under the multi-modal setting (+3.54 mIoU) and excels in the challenging modality-incomplete setting(+19.79 mIoU).
Authors: Muyang Qiu, Jian Zhang, Lei Qi, Qian Yu, Yinghuan Shi, Yang Gao
Abstract: Despite the recent success of domain generalization in medical image segmentation, voxel-wise annotation for all source domains remains a huge burden. Semi-supervised domain generalization has been proposed very recently to combat this challenge by leveraging limited labeled data along with abundant unlabeled data collected from multiple medical institutions, depending on precisely harnessing unlabeled data while improving generalization simultaneously. In this work, we observe that domain shifts between medical institutions cause disparate feature statistics, which significantly deteriorates pseudo-label quality due to an unexpected normalization process. Nevertheless, this phenomenon could be exploited to facilitate unseen domain generalization. Therefore, we propose 1) multiple statistics-individual branches to mitigate the impact of domain shifts for reliable pseudo-labels and 2) one statistics-aggregated branch for domain-invariant feature learning. Furthermore, to simulate unseen domains with statistics difference, we approach this from two aspects, i.e., a perturbation with histogram matching at image level and a random batch normalization selection strategy at feature level, producing diverse statistics to expand the training distribution. Evaluation results on three medical image datasets demonstrate the effectiveness of our method compared with recent SOTA methods. The code is available at https://github.com/qiumuyang/SIAB.
Authors: Hyeon Bae Kim, Yong Hyun Ahn, Seong Tae Kim
Abstract: Recent advancements in deep neural networks have shown promise in aiding disease diagnosis and medical decision-making. However, ensuring transparent decision-making processes of AI models in compliance with regulations requires a comprehensive understanding of the model's internal workings. However, previous methods heavily rely on expensive pixel-wise annotated datasets for interpreting the model, presenting a significant drawback in medical domains. In this paper, we propose a novel medical neuron concept annotation method, named Mask-free Medical Model Interpretation (MAMMI), addresses these challenges. By using a vision-language model, our method relaxes the need for pixel-level masks for neuron concept annotation. MAMMI achieves superior performance compared to other interpretation methods, demonstrating its efficacy in providing rich representations for neurons in medical image analysis. Our experiments on a model trained on NIH chest X-rays validate the effectiveness of MAMMI, showcasing its potential for transparent clinical decision-making in the medical domain. The code is available at https://github.com/ailab-kyunghee/MAMMI.
Authors: Yucheng Lu, Dovile Juodelyte, Jonathan Victor, Veronika Cheplygina
Abstract: In this paper, we use spectral analysis to investigate transfer learning and study model sensitivity to frequency shortcuts in medical imaging. By analyzing the power spectrum density of both pre-trained and fine-tuned model gradients, as well as artificially generated frequency shortcuts, we observe notable differences in learning priorities between models pre-trained on natural vs medical images, which generally persist during fine-tuning. We find that when a model's learning priority aligns with the power spectrum density of an artifact, it results in overfitting to that artifact. Based on these observations, we show that source data editing can alter the model's resistance to shortcut learning.
Authors: Chenyu Liu, Jia Pan, Jinshui Hu, Baocai Yin, Bing Yin, Mingjun Chen, Cong Liu, Jun Du, Qingfeng Liu
Abstract: Recently, Handwritten Mathematical Expression Recognition (HMER) has gained considerable attention in pattern recognition for its diverse applications in document understanding. Current methods typically approach HMER as an image-to-sequence generation task within an autoregressive (AR) encoder-decoder framework. However, these approaches suffer from several drawbacks: 1) a lack of overall language context, limiting information utilization beyond the current decoding step; 2) error accumulation during AR decoding; and 3) slow decoding speed. To tackle these problems, this paper makes the first attempt to build a novel bottom-up Non-AutoRegressive Modeling approach for HMER, called NAMER. NAMER comprises a Visual Aware Tokenizer (VAT) and a Parallel Graph Decoder (PGD). Initially, the VAT tokenizes visible symbols and local relations at a coarse level. Subsequently, the PGD refines all tokens and establishes connectivities in parallel, leveraging comprehensive visual and linguistic contexts. Experiments on CROHME 2014/2016/2019 and HME100K datasets demonstrate that NAMER not only outperforms the current state-of-the-art (SOTA) methods on ExpRate by 1.93%/2.35%/1.49%/0.62%, but also achieves significant speedups of 13.7x and 6.7x faster in decoding time and overall FPS, proving the effectiveness and efficiency of NAMER.
Authors: Yunya Gao
Abstract: Updated building footprints with refugee camps from high-resolution satellite imagery can support related humanitarian operations. This study explores the utilization of the "Segment Anything Model" (SAM) and one of its branches, SAM-Adapter, for semantic segmentation tasks in the building extraction from satellite imagery. SAM-Adapter is a lightweight adaptation of the SAM and emerges as a powerful tool for this extraction task across diverse refugee camps. Our research proves that SAM-Adapter excels in scenarios where data availability is limited compared to other classic (e.g., U-Net) or advanced semantic segmentation models (e.g., Transformer). Furthermore, the impact of upscaling techniques on model performance is highlighted, with methods like super-resolution (SR) models proving invaluable for improving model performance. Additionally, the study unveils intriguing phenomena, including the model's rapid convergence in the first training epoch when using upscaled image data for training, suggesting opportunities for future research. The codes covering each step from data preparation, model training, model inferencing, and the generation of Shapefiles for predicted masks are available on a GitHub repository to benefit the extended scientific community and humanitarian operations.
Authors: Jianhao Li, Tianyu Sun, Zhongdao Wang, Enze Xie, Bailan Feng, Hongbo Zhang, Ze Yuan, Ke Xu, Jiaheng Liu, Ping Luo
Abstract: This paper proposes an algorithm for automatically labeling 3D objects from 2D point or box prompts, especially focusing on applications in autonomous driving. Unlike previous arts, our auto-labeler predicts 3D shapes instead of bounding boxes and does not require training on a specific dataset. We propose a Segment, Lift, and Fit (SLF) paradigm to achieve this goal. Firstly, we segment high-quality instance masks from the prompts using the Segment Anything Model (SAM) and transform the remaining problem into predicting 3D shapes from given 2D masks. Due to the ill-posed nature of this problem, it presents a significant challenge as multiple 3D shapes can project into an identical mask. To tackle this issue, we then lift 2D masks to 3D forms and employ gradient descent to adjust their poses and shapes until the projections fit the masks and the surfaces conform to surrounding LiDAR points. Notably, since we do not train on a specific dataset, the SLF auto-labeler does not overfit to biased annotation patterns in the training set as other methods do. Thus, the generalization ability across different datasets improves. Experimental results on the KITTI dataset demonstrate that the SLF auto-labeler produces high-quality bounding box annotations, achieving an AP@0.5 IoU of nearly 90\%. Detectors trained with the generated pseudo-labels perform nearly as well as those trained with actual ground-truth annotations. Furthermore, the SLF auto-labeler shows promising results in detailed shape predictions, providing a potential alternative for the occupancy annotation of dynamic objects.
Authors: Tonmoy Rajkhowa, Amartya Roy Chowdhury, Sankalp Nagaonkar, Achyut Mani Tripathi
Abstract: In healthcare and medical diagnostics, Visual Question Answering (VQA) mayemergeasapivotal tool in scenarios where analysis of intricate medical images becomes critical for accurate diagnoses. Current text-based VQA systems limit their utility in scenarios where hands-free interaction and accessibility are crucial while performing tasks. A speech-based VQA system may provide a better means of interaction where information can be accessed while performing tasks simultaneously. To this end, this work implements a speech-based VQA system by introducing a Textless Multilingual Pathological VQA (TMPathVQA) dataset, an expansion of the PathVQA dataset, containing spoken questions in English, German & French. This dataset comprises 98,397 multilingual spoken questions and answers based on 5,004 pathological images along with 70 hours of audio. Finally, this work benchmarks and compares TMPathVQA systems implemented using various combinations of acoustic and visual features.
Authors: Kalliopi Basioti, Mohamed A. Abdelsalam, Federico Fancellu, Vladimir Pavlovic, Afsaneh Fazly
Abstract: Controllable Image Captioning (CIC) aims at generating natural language descriptions for an image, conditioned on information provided by end users, e.g., regions, entities or events of interest. However, available image--language datasets mainly contain captions that describe the entirety of an image, making them ineffective for training CIC models that can potentially attend to any subset of regions or relationships. To tackle this challenge, we propose a novel, fully automatic method to sample additional focused and visually grounded captions using a unified structured semantic representation built on top of the existing set of captions associated with an image. We leverage Abstract Meaning Representation (AMR), a cross-lingual graph-based semantic formalism, to encode all possible spatio-semantic relations between entities, beyond the typical spatial-relations-only focus of current methods. We use this Structured Semantic Augmentation (SSA) framework to augment existing image--caption datasets with the grounded controlled captions, increasing their spatial and semantic diversity and focal coverage. We then develop a new model, CIC-BART-SSA, specifically tailored for the CIC task, that sources its control signals from SSA-diversified datasets. We empirically show that, compared to SOTA CIC models, CIC-BART-SSA generates captions that are superior in diversity and text quality, are competitive in controllability, and, importantly, minimize the gap between broad and highly focused controlled captioning performance by efficiently generalizing to the challenging highly focused scenarios. Code is available at https://github.com/SamsungLabs/CIC-BART-SSA.
Authors: Jiwook Kim, Seonho Lee, Jaeyo Shin, Jiho Choi, Hyunjung Shim
Abstract: Score distillation sampling (SDS) has emerged as an effective framework in text-driven 3D editing tasks due to its inherent 3D consistency. However, existing SDS-based 3D editing methods suffer from extensive training time and lead to low-quality results, primarily because these methods deviate from the sampling dynamics of diffusion models. In this paper, we propose DreamCatalyst, a novel framework that interprets SDS-based editing as a diffusion reverse process. Our objective function considers the sampling dynamics, thereby making the optimization process of DreamCatalyst an approximation of the diffusion reverse process in editing tasks. DreamCatalyst aims to reduce training time and improve editing quality. DreamCatalyst presents two modes: (1) a faster mode, which edits the NeRF scene in only about 25 minutes, and (2) a high-quality mode, which produces superior results in less than 70 minutes. Specifically, our high-quality mode outperforms current state-of-the-art NeRF editing methods both in terms of speed and quality. See more extensive results on our project page: https://dream-catalyst.github.io.
Authors: Yanqin Jiang, Chaohui Yu, Chenjie Cao, Fan Wang, Weiming Hu, Jin Gao
Abstract: Recent advances in 4D generation mainly focus on generating 4D content by distilling pre-trained text or single-view image-conditioned models. It is inconvenient for them to take advantage of various off-the-shelf 3D assets with multi-view attributes, and their results suffer from spatiotemporal inconsistency owing to the inherent ambiguity in the supervision signals. In this work, we present Animate3D, a novel framework for animating any static 3D model. The core idea is two-fold: 1) We propose a novel multi-view video diffusion model (MV-VDM) conditioned on multi-view renderings of the static 3D object, which is trained on our presented large-scale multi-view video dataset (MV-Video). 2) Based on MV-VDM, we introduce a framework combining reconstruction and 4D Score Distillation Sampling (4D-SDS) to leverage the multi-view video diffusion priors for animating 3D objects. Specifically, for MV-VDM, we design a new spatiotemporal attention module to enhance spatial and temporal consistency by integrating 3D and video diffusion models. Additionally, we leverage the static 3D model's multi-view renderings as conditions to preserve its identity. For animating 3D models, an effective two-stage pipeline is proposed: we first reconstruct motions directly from generated multi-view videos, followed by the introduced 4D-SDS to refine both appearance and motion. Qualitative and quantitative experiments demonstrate that Animate3D significantly outperforms previous approaches. Data, code, and models will be open-released.
Authors: Ruijie Yang, Yan Zhu, Peiyao Fu, Yizhe Zhang, Zhihua Wang, Quanlin Li, Pinghong Zhou, Xian Yang, Shuo Wang
Abstract: Determining the necessity of resecting malignant polyps during colonoscopy screen is crucial for patient outcomes, yet challenging due to the time-consuming and costly nature of histopathology examination. While deep learning-based classification models have shown promise in achieving optical biopsy with endoscopic images, they often suffer from a lack of explainability. To overcome this limitation, we introduce EndoFinder, a content-based image retrieval framework to find the 'digital twin' polyp in the reference database given a newly detected polyp. The clinical semantics of the new polyp can be inferred referring to the matched ones. EndoFinder pioneers a polyp-aware image encoder that is pre-trained on a large polyp dataset in a self-supervised way, merging masked image modeling with contrastive learning. This results in a generic embedding space ready for different downstream clinical tasks based on image retrieval. We validate the framework on polyp re-identification and optical biopsy tasks, with extensive experiments demonstrating that EndoFinder not only achieves explainable diagnostics but also matches the performance of supervised classification models. EndoFinder's reliance on image retrieval has the potential to support diverse downstream decision-making tasks during real-time colonoscopy procedures.
Authors: Yang Zhou, Yongjian Wu, Jiya Saiyin, Bingzheng Wei, Maode Lai, Eric Chang, Yan Xu
Abstract: Prompt tuning methods have achieved remarkable success in parameter-efficient fine-tuning on large pre-trained models. However, their application to dual-modal fusion-based visual-language pre-trained models (VLPMs), such as GLIP, has encountered issues. Existing prompt tuning methods have not effectively addressed the modal mapping and aligning problem for tokens in different modalities, leading to poor transfer generalization. To address this issue, we propose Synchronous Dual Prompt Tuning (SDPT). SDPT initializes a single set of learnable unified prototype tokens in the established modal aligning space to represent the aligned semantics of text and image modalities for downstream tasks. Furthermore, SDPT establishes inverse linear projections that require no training to embed the information of unified prototype tokens into the input space of different modalities. The inverse linear projections allow the unified prototype token to synchronously represent the two modalities and enable SDPT to share the unified semantics of text and image for downstream tasks across different modal prompts. Experimental results demonstrate that SDPT assists fusion-based VLPMs to achieve superior outcomes with only 0.04\% of model parameters for training across various scenarios, outperforming other single- or dual-modal methods. The code will be released at https://github.com/wuyongjianCODE/SDPT.
Authors: Chenfan Xu, Zhentao Liu, Yuan Liu, Yulong Dou, Jiamin Wu, Jiepeng Wang, Minjiao Wang, Dinggang Shen, Zhiming Cui
Abstract: Orthodontic treatment usually requires regular face-to-face examinations to monitor dental conditions of the patients. When in-person diagnosis is not feasible, an alternative is to utilize five intra-oral photographs for remote dental monitoring. However, it lacks of 3D information, and how to reconstruct 3D dental models from such sparse view photographs is a challenging problem. In this study, we propose a 3D teeth reconstruction framework, named TeethDreamer, aiming to restore the shape and position of the upper and lower teeth. Given five intra-oral photographs, our approach first leverages a large diffusion model's prior knowledge to generate novel multi-view images with known poses to address sparse inputs and then reconstructs high-quality 3D teeth models by neural surface reconstruction. To ensure the 3D consistency across generated views, we integrate a 3D-aware feature attention mechanism in the reverse diffusion process. Moreover, a geometry-aware normal loss is incorporated into the teeth reconstruction process to enhance geometry accuracy. Extensive experiments demonstrate the superiority of our method over current state-of-the-arts, giving the potential to monitor orthodontic treatment remotely. Our code is available at https://github.com/ShanghaiTech-IMPACT/TeethDreamer
Authors: Jinrui Zhang, Teng Wang, Haigang Zhang, Ping Lu, Feng Zheng
Abstract: Large vision-language models (LVLMs) have shown promising performance on a variety of vision-language tasks. However, they remain susceptible to hallucinations, generating outputs misaligned with visual content or instructions. While various mitigation strategies have been proposed, they often neglect a key contributor to hallucinations: lack of fine-grained reasoning supervision during training. Without intermediate reasoning steps, models may establish superficial shortcuts between instructions and responses, failing to internalize the inherent reasoning logic. To address this challenge, we propose reflective instruction tuning, which integrates rationale learning into visual instruction tuning. Unlike previous methods that learning from responses only, our approach entails the model predicting rationales justifying why responses are correct or incorrect. This fosters a deeper engagement with the fine-grained reasoning underlying each response, thus enhancing the model's reasoning proficiency. To facilitate this approach, we propose REVERIE, the first large-scale instruction-tuning dataset with ReflEctiVE RatIonalE annotations. REVERIE comprises 115k machine-generated reasoning instructions, each meticulously annotated with a corresponding pair of correct and confusing responses, alongside comprehensive rationales elucidating the justification behind the correctness or erroneousness of each response. Experimental results on multiple LVLM benchmarks reveal that reflective instruction tuning with the REVERIE dataset yields noticeable performance gain over the baseline model, demonstrating the effectiveness of reflecting from the rationales. Project page is at https://zjr2000.github.io/projects/reverie.
Authors: Ouxiang Li, Yanbin Hao, Zhicai Wang, Bin Zhu, Shuo Wang, Zaixi Zhang, Fuli Feng
Abstract: Model inversion attacks (MIAs) aim to reconstruct private images from a target classifier's training set, thereby raising privacy concerns in AI applications. Previous GAN-based MIAs tend to suffer from inferior generative fidelity due to GAN's inherent flaws and biased optimization within latent space. To alleviate these issues, leveraging on diffusion models' remarkable synthesis capabilities, we propose Diffusion-based Model Inversion (Diff-MI) attacks. Specifically, we introduce a novel target-specific conditional diffusion model (CDM) to purposely approximate target classifier's private distribution and achieve superior accuracy-fidelity balance. Our method involves a two-step learning paradigm. Step-1 incorporates the target classifier into the entire CDM learning under a pretrain-then-finetune fashion, with creating pseudo-labels as model conditions in pretraining and adjusting specified layers with image predictions in fine-tuning. Step-2 presents an iterative image reconstruction method, further enhancing the attack performance through a combination of diffusion priors and target knowledge. Additionally, we propose an improved max-margin loss that replaces the hard max with top-k maxes, fully leveraging feature information and soft labels from the target classifier. Extensive experiments demonstrate that Diff-MI significantly improves generative fidelity with an average decrease of 20% in FID while maintaining competitive attack accuracy compared to state-of-the-art methods across various datasets and models. We will release our code and models.
Authors: Chang Li, Jiao Guo, Yufei Zhao, Yongjun Zhang
Abstract: This paper is the first to propose an end-to-end framework of mutually reinforcing images to 3D surface recurrent neural network-like for model-adaptation indoor 3D reconstruction,where multi-view dense matching and point cloud surface optimization are mutually reinforced by a RNN-like structure rather than being treated as a separate issue.The characteristics are as follows:In the multi-view dense matching module, the model-adaptation strategy is used to fine-tune and optimize a Transformer-based multi-view dense matching DNN,so that it has the higher image feature for matching and detail expression capabilities;In the point cloud surface optimization module,the 3D surface reconstruction network based on 3D implicit field is optimized by using model-adaptation strategy,which solves the problem of point cloud surface optimization without knowing normal vector of 3D surface.To improve and finely reconstruct 3D surfaces from point cloud,smooth loss is proposed and added to this module;The MRIo3DS-Net is a RNN-like framework,which utilizes the finely optimized 3D surface obtained by PCSOM to recursively reinforce the differentiable warping for optimizing MVDMM.This refinement leads to achieving better dense matching results, and better dense matching results leads to achieving better 3D surface results recursively and mutually.Hence, model-adaptation strategy can better collaborate the differences between the two network modules,so that they complement each other to achieve the better effect;To accelerate the transfer learning and training convergence from source domain to target domain,a multi-task loss function based on Bayesian uncertainty is used to adaptively adjust the weights between the two networks loss functions of MVDMM and PCSOM;In this multi-task cascade network framework,any modules can be replaced by any state-of-the-art networks to achieve better 3D reconstruction results.
Authors: Yisen Wang, Yao Teng, Limin Wang
Abstract: Recognition and generation are two fundamental tasks in computer vision, which are often investigated separately in the exiting literature. However, these two tasks are highly correlated in essence as they both require understanding the underline semantics of visual concepts. In this paper, we propose a new learning framework, coined as CycleHOI, to boost the performance of human-object interaction (HOI) detection by bridging the DETR-based detection pipeline and the pre-trained text-to-image diffusion model. Our key design is to introduce a novel cycle consistency loss for the training of HOI detector, which is able to explicitly leverage the knowledge captured in the powerful diffusion model to guide the HOI detector training. Specifically, we build an extra generation task on top of the decoded instance representations from HOI detector to enforce a detection-generation cycle consistency. Moreover, we perform feature distillation from diffusion model to detector encoder to enhance its representation power. In addition, we further utilize the generation power of diffusion model to augment the training set in both aspects of label correction and sample generation. We perform extensive experiments to verify the effectiveness and generalization power of our CycleHOI with three HOI detection frameworks on two public datasets: HICO-DET and V-COCO. The experimental results demonstrate our CycleHOI can significantly improve the performance of the state-of-the-art HOI detectors.
Authors: Yihang Chen, Tsai Hor Chan, Guosheng Yin, Yuming Jiang, Lequan Yu
Abstract: Multiple instance learning (MIL) has been extensively applied to whole slide histopathology image (WSI) analysis. The existing aggregation strategy in MIL, which primarily relies on the first-order distance (e.g., mean difference) between instances, fails to accurately approximate the true feature distribution of each instance, leading to biased slide-level representations. Moreover, the scarcity of WSI observations easily leads to model overfitting, resulting in unstable testing performance and limited generalizability. To tackle these challenges, we propose a new Bayesian nonparametric framework for multiple instance learning, which adopts a cascade of Dirichlet processes (cDP) to incorporate the instance-to-bag characteristic of the WSIs. We perform feature aggregation based on the latent clusters formed by the Dirichlet process, which incorporates the covariances of the patch features and forms more representative clusters. We then perform bag-level prediction with another Dirichlet process model on the bags, which imposes a natural regularization on learning to prevent overfitting and enhance generalizability. Moreover, as a Bayesian nonparametric method, the cDP model can accurately generate posterior uncertainty, which allows for the detection of outlier samples and tumor localization. Extensive experiments on five WSI benchmarks validate the superior performance of our method, as well as its generalizability and ability to estimate uncertainties. Codes are available at https://github.com/HKU-MedAI/cDPMIL.
Authors: Shunqi Mao, Chaoyi Zhang, Hang Su, Hwanjun Song, Igor Shalyminov, Weidong Cai
Abstract: Contextualized Image Captioning (CIC) evolves traditional image captioning into a more complex domain, necessitating the ability for multimodal reasoning. It aims to generate image captions given specific contextual information. This paper further introduces a novel domain of Controllable Contextualized Image Captioning (Ctrl-CIC). Unlike CIC, which solely relies on broad context, Ctrl-CIC accentuates a user-defined highlight, compelling the model to tailor captions that resonate with the highlighted aspects of the context. We present two approaches, Prompting-based Controller (P-Ctrl) and Recalibration-based Controller (R-Ctrl), to generate focused captions. P-Ctrl conditions the model generation on highlight by prepending captions with highlight-driven prefixes, whereas R-Ctrl tunes the model to selectively recalibrate the encoder embeddings for highlighted tokens. Additionally, we design a GPT-4V empowered evaluator to assess the quality of the controlled captions alongside standard assessment methods. Extensive experimental results demonstrate the efficient and effective controllability of our method, charting a new direction in achieving user-adaptive image captioning. Code is available at https://github.com/ShunqiM/Ctrl-CIC .
Authors: Zhi Cai, Yingjie Gao, Yaoyan Zheng, Nan Zhou, Di Huang
Abstract: In computer vision, object detection is an important task that finds its application in many scenarios. However, obtaining extensive labels can be challenging, especially in crowded scenes. Recently, the Segment Anything Model (SAM) has been proposed as a powerful zero-shot segmenter, offering a novel approach to instance segmentation tasks. However, the accuracy and efficiency of SAM and its variants are often compromised when handling objects in crowded and occluded scenes. In this paper, we introduce Crowd-SAM, a SAM-based framework designed to enhance SAM's performance in crowded and occluded scenes with the cost of few learnable parameters and minimal labeled images. We introduce an efficient prompt sampler (EPS) and a part-whole discrimination network (PWD-Net), enhancing mask selection and accuracy in crowded scenes. Despite its simplicity, Crowd-SAM rivals state-of-the-art (SOTA) fully-supervised object detection methods on several benchmarks including CrowdHuman and CityPersons. Our code is available at https://github.com/FelixCaae/CrowdSAM.
Authors: Qiaoqiao Jin, Rui Shi, Yishun Dou, Bingbing Ni
Abstract: Current Facial Action Unit (FAU) detection methods generally encounter difficulties due to the scarcity of labeled video training data and the limited number of training face IDs, which renders the trained feature extractor insufficient coverage for modeling the large diversity of inter-person facial structures and movements. To explicitly address the above challenges, we propose a novel video-level pre-training scheme by fully exploring the multi-label property of FAUs in the video as well as the temporal label consistency. At the heart of our design is a pre-trained video feature extractor based on the video-masked autoencoder together with a fine-tuning network that jointly completes the multi-level video FAUs analysis tasks, \emph{i.e.} integrating both video-level and frame-level FAU detections, thus dramatically expanding the supervision set from sparse FAUs annotations to ALL video frames including masked ones. Moreover, we utilize inter-frame and intra-frame AU pair state matrices as prior knowledge to guide network training instead of traditional Graph Neural Networks, for better temporal supervision. Our approach demonstrates substantial enhancement in performance compared to the existing state-of-the-art methods used in BP4D and DISFA FAUs datasets.
Authors: Jialong Huang, Gaojie Li, Shichao Kan, Jianfeng Liu, Yixiong Liang
Abstract: Current cervical cytopathology whole slide image (WSI) screening primarily relies on detection-based approaches, which are limited in performance due to the expense and time-consuming annotation process. Multiple Instance Learning (MIL), a weakly supervised approach that relies solely on bag-level labels, can effectively alleviate these challenges. Nonetheless, MIL commonly employs frozen pretrained models or self-supervised learning for feature extraction, which suffers from low efficacy or inefficiency. In this paper, we propose an efficient framework for cervical cytopathology WSI classification using only WSI-level labels through unsupervised and weakly supervised learning. Given the sparse and dispersed nature of abnormal cells within cytopathological WSIs, we propose a strategy that leverages the pretrained foundation model to filter the top$k$ high-risk patches. Subsequently, we suggest parameter-efficient fine-tuning (PEFT) of a large foundation model using contrastive learning on the filtered patches to enhance its representation ability for task-specific signals. By training only the added linear adapters, we enhance the learning of patch-level features with substantially reduced time and memory consumption. Experiments conducted on the CSD and FNAC 2019 datasets demonstrate that the proposed method enhances the performance of various MIL methods and achieves state-of-the-art (SOTA) performance. The code and trained models are publicly available at https://github.com/CVIU-CSU/TCT-InfoNCE.
Authors: Renjie Lu, Jingke Meng, Wei-Shi Zheng
Abstract: Vision and language navigation is a task that requires an agent to navigate according to a natural language instruction. Recent methods predict sub-goals on constructed topology map at each step to enable long-term action planning. However, they suffer from high computational cost when attempting to support such high-level predictions with GCN-like models. In this work, we propose an alternative method that facilitates navigation planning by considering the alignment between instructions and directed fidelity trajectories, which refers to a path from the initial node to the candidate locations on a directed graph without detours. This planning strategy leads to an efficient model while achieving strong performance. Specifically, we introduce a directed graph to illustrate the explored area of the environment, emphasizing directionality. Then, we firstly define the trajectory representation as a sequence of directed edge features, which are extracted from the panorama based on the corresponding orientation. Ultimately, we assess and compare the alignment between instruction and different trajectories during navigation to determine the next navigation target. Our method outperforms previous SOTA method BEVBert on RxR dataset and is comparable on R2R dataset while largely reducing the computational cost. Code is available: https://github.com/iSEE-Laboratory/VLN-PRET.
Authors: Guowei Xu, Jiale Tao, Wen Li, Lixin Duan
Abstract: In the realm of stochastic human motion prediction (SHMP), researchers have often turned to generative models like GANS, VAEs and diffusion models. However, most previous approaches have struggled to accurately predict motions that are both realistic and coherent with past motion due to a lack of guidance on the latent distribution. In this paper, we introduce Semantic Latent Directions (SLD) as a solution to this challenge, aiming to constrain the latent space to learn meaningful motion semantics and enhance the accuracy of SHMP. SLD defines a series of orthogonal latent directions and represents the hypothesis of future motion as a linear combination of these directions. By creating such an information bottleneck, SLD excels in capturing meaningful motion semantics, thereby improving the precision of motion predictions. Moreover, SLD offers controllable prediction capabilities by adjusting the coefficients of the latent directions during the inference phase. Expanding on SLD, we introduce a set of motion queries to enhance the diversity of predictions. By aligning these motion queries with the SLD space, SLD is further promoted to more accurate and coherent motion predictions. Through extensive experiments conducted on widely used benchmarks, we showcase the superiority of our method in accurately predicting motions while maintaining a balance of realism and diversity. Our code and pretrained models are available at https://github.com/GuoweiXu368/SLD-HMP.
Authors: Qijie Mo, Yipeng Gao, Shenghao Fu, Junkai Yan, Ancong Wu, Wei-Shi Zheng
Abstract: In incremental object detection, knowledge distillation has been proven to be an effective way to alleviate catastrophic forgetting. However, previous works focused on preserving the knowledge of old models, ignoring that images could simultaneously contain categories from past, present, and future stages. The co-occurrence of objects makes the optimization objectives inconsistent across different stages since the definition for foreground objects differs across various stages, which limits the model's performance greatly. To overcome this problem, we propose a method called ``Bridge Past and Future'' (BPF), which aligns models across stages, ensuring consistent optimization directions. In addition, we propose a novel Distillation with Future (DwF) loss, fully leveraging the background probability to mitigate the forgetting of old classes while ensuring a high level of adaptability in learning new classes. Extensive experiments are conducted on both Pascal VOC and MS COCO benchmarks. Without memory, BPF outperforms current state-of-the-art methods under various settings. The code is available at https://github.com/iSEE-Laboratory/BPF.
Authors: Boqiang Zhang, Zuan Gao, Yadong Qu, Hongtao Xie
Abstract: Visual text generation has significantly advanced through diffusion models aimed at producing images with readable and realistic text. Recent works primarily use a ControlNet-based framework, employing standard font text images to control diffusion models. Recognizing the critical role of control information in generating high-quality text, we investigate its influence from three perspectives: input encoding, role at different stages, and output features. Our findings reveal that: 1) Input control information has unique characteristics compared to conventional inputs like Canny edges and depth maps. 2) Control information plays distinct roles at different stages of the denoising process. 3) Output control features significantly differ from the base and skip features of the U-Net decoder in the frequency domain. Based on these insights, we propose TextGen, a novel framework designed to enhance generation quality by optimizing control information. We improve input and output features using Fourier analysis to emphasize relevant information and reduce noise. Additionally, we employ a two-stage generation framework to align the different roles of control information at different stages. Furthermore, we introduce an effective and lightweight dataset for training. Our method achieves state-of-the-art performance in both Chinese and English text generation. The code and dataset will be made available.
Authors: Shijie Chang, Youwei Pang, Xiaoqi Zhao, Lihe Zhang, Huchuan Lu
Abstract: Existing few-shot segmentation (FSS) methods mainly focus on prototype feature generation and the query-support matching mechanism. As a crucial prompt for generating prototype features, the pair of image-mask types in the support set has become the default setting. However, various types such as image, text, box, and mask all can provide valuable information regarding the objects in context, class, localization, and shape appearance. Existing work focuses on specific combinations of guidance, leading FSS into different research branches. Rethinking guidance types in FSS is expected to explore the efficient joint representation of the coupling between the support set and query set, giving rise to research trends in the weakly or strongly annotated guidance to meet the customized requirements of practical users. In this work, we provide the generalized FSS with seven guidance paradigms and develop a universal vision-language framework (UniFSS) to integrate prompts from text, mask, box, and image. Leveraging the advantages of large-scale pre-training vision-language models in textual and visual embeddings, UniFSS proposes high-level spatial correction and embedding interactive units to overcome the semantic ambiguity drawbacks typically encountered by pure visual matching methods when facing intra-class appearance diversities. Extensive experiments show that UniFSS significantly outperforms the state-of-the-art methods. Notably, the weakly annotated class-aware box paradigm even surpasses the finely annotated mask paradigm.
Authors: Lihan Tong, Yun Liu, Weijia Li, Liyuan Chen, Erkang Chen
Abstract: Single-image dehazing is a pivotal challenge in computer vision that seeks to remove haze from images and restore clean background details. Recognizing the limitations of traditional physical model-based methods and the inefficiencies of current attention-based solutions, we propose a new dehazing network combining an innovative Haze-Aware Attention Module (HAAM) with a Multiscale Frequency Enhancement Module (MFEM). The HAAM is inspired by the atmospheric scattering model, thus skillfully integrating physical principles into high-dimensional features for targeted dehazing. It picks up on latent features during the image restoration process, which gives a significant boost to the metrics, while the MFEM efficiently enhances high-frequency details, thus sidestepping wavelet or Fourier transform complexities. It employs multiscale fields to extract and emphasize key frequency components with minimal parameter overhead. Integrated into a simple U-Net framework, our Haze-Aware Attention Network (HAA-Net) for single-image dehazing significantly outperforms existing attention-based and transformer models in efficiency and effectiveness. Tested across various public datasets, the HAA-Net sets new performance benchmarks. Our work not only advances the field of image dehazing but also offers insights into the design of attention mechanisms for broader applications in computer vision.
Authors: Ludovica Schaerf, Andrea Alfarano, Eric Postma
Abstract: Colorway creation is the task of generating textile samples in alternate color variations maintaining an underlying pattern. The individuation of a suitable color palette for a colorway is a complex creative task, responding to client and market needs, stylistic and cultural specifications, and mood. We introduce a modification of this task, the "generative colorway" creation, that includes minimal shape modifications, and propose a framework, "ColorwAI", to tackle this task using color disentanglement on StyleGAN and Diffusion. We introduce a variation of the InterfaceGAN method for supervised disentanglement, ShapleyVec. We use Shapley values to subselect a few dimensions of the detected latent direction. Moreover, we introduce a general framework to adopt common disentanglement methods on any architecture with a semantic latent space and test it on Diffusion and GANs. We interpret the color representations within the models' latent space. We find StyleGAN's W space to be the most aligned with human notions of color. Finally, we suggest that disentanglement can solicit a creative system for colorway creation, and evaluate it through expert questionnaires and creativity theory.
Authors: Pengxiang Li, Zhi Gao, Bofei Zhang, Tao Yuan, Yuwei Wu, Mehrtash Harandi, Yunde Jia, Song-Chun Zhu, Qing Li
Abstract: Vision language models (VLMs) have achieved impressive progress in diverse applications, becoming a prevalent research direction. In this paper, we build FIRE, a feedback-refinement dataset, consisting of 1.1M multi-turn conversations that are derived from 27 source datasets, empowering VLMs to spontaneously refine their responses based on user feedback across diverse tasks. To scale up the data collection, FIRE is collected in two components: FIRE-100K and FIRE-1M, where FIRE-100K is generated by GPT-4V, and FIRE-1M is freely generated via models trained on FIRE-100K. Then, we build FIRE-Bench, a benchmark to comprehensively evaluate the feedback-refining capability of VLMs, which contains 11K feedback-refinement conversations as the test data, two evaluation settings, and a model to provide feedback for VLMs. We develop the FIRE-LLaVA model by fine-tuning LLaVA on FIRE-100K and FIRE-1M, which shows remarkable feedback-refining capability on FIRE-Bench and outperforms untrained VLMs by 50%, making more efficient user-agent interactions and underscoring the significance of the FIRE dataset.
Authors: Alessio Sampieri, Alessio Palma, Indro Spinelli, Fabio Galasso
Abstract: The target duration of a synthesized human motion is a critical attribute that requires modeling control over the motion dynamics and style. Speeding up an action performance is not merely fast-forwarding it. However, state-of-the-art techniques for human behavior synthesis have limited control over the target sequence length. We introduce the problem of generating length-aware 3D human motion sequences from textual descriptors, and we propose a novel model to synthesize motions of variable target lengths, which we dub "Length-Aware Latent Diffusion" (LADiff). LADiff consists of two new modules: 1) a length-aware variational auto-encoder to learn motion representations with length-dependent latent codes; 2) a length-conforming latent diffusion model to generate motions with a richness of details that increases with the required target sequence length. LADiff significantly improves over the state-of-the-art across most of the existing motion synthesis metrics on the two established benchmarks of HumanML3D and KIT-ML.
Authors: Wenzhao Xiang, Chang Liu, Hang Su, Hongyang Yu
Abstract: Masked image modeling (MIM) has gained significant traction for its remarkable prowess in representation learning. As an alternative to the traditional approach, the reconstruction from corrupted images has recently emerged as a promising pretext task. However, the regular corrupted images are generated using generic generators, often lacking relevance to the specific reconstruction task involved in pre-training. Hence, reconstruction from regular corrupted images cannot ensure the difficulty of the pretext task, potentially leading to a performance decline. Moreover, generating corrupted images might introduce an extra generator, resulting in a notable computational burden. To address these issues, we propose to incorporate adversarial examples into masked image modeling, as the new reconstruction targets. Adversarial examples, generated online using only the trained models, can directly aim to disrupt tasks associated with pre-training. Therefore, the incorporation not only elevates the level of challenge in reconstruction but also enhances efficiency, contributing to the acquisition of superior representations by the model. In particular, we introduce a novel auxiliary pretext task that reconstructs the adversarial examples corresponding to the original images. We also devise an innovative adversarial attack to craft more suitable adversarial examples for MIM pre-training. It is noted that our method is not restricted to specific model architectures and MIM strategies, rendering it an adaptable plug-in capable of enhancing all MIM methods. Experimental findings substantiate the remarkable capability of our approach in amplifying the generalization and robustness of existing MIM methods. Notably, our method surpasses the performance of baselines on various tasks, including ImageNet, its variants, and other downstream tasks.
Authors: Hyunchul Bae, Minhee Kang, Heejin Ahn
Abstract: In this paper, we investigate improving the perception performance of autonomous vehicles through communication with other vehicles and road infrastructures. To this end, we introduce a collaborative perception model $\textbf{V2X-M2C}$, consisting of multiple modules, each generating inter-agent complementary information, spatial global context, and spatial local information. Inspired by the question of why most existing architectures are sequential, we analyze both the $\textit{sequential}$ and $\textit{parallel}$ connections of the modules. The sequential connection synergizes the modules, whereas the parallel connection independently improves each module. Extensive experiments demonstrate that V2X-M2C achieves state-of-the-art perception performance, increasing the detection accuracy by 8.00% to 10.87% and decreasing the FLOPs by 42.81% to 52.64%.
Authors: Soobin Um, Jong Chul Ye
Abstract: We present a novel approach for generating minority samples that live on low-density regions of a data manifold. Our framework is built upon diffusion models, leveraging the principle of guided sampling that incorporates an arbitrary energy-based guidance during inference time. The key defining feature of our sampler lies in its \emph{self-contained} nature, \ie, implementable solely with a pretrained model. This distinguishes our sampler from existing techniques that require expensive additional components (like external classifiers) for minority generation. Specifically, we first estimate the likelihood of features within an intermediate latent sample by evaluating a reconstruction loss w.r.t. its posterior mean. The generation then proceeds with the minimization of the estimated likelihood, thereby encouraging the emergence of minority features in the latent samples of subsequent timesteps. To further improve the performance of our sampler, we provide several time-scheduling techniques that properly manage the influence of guidance over inference steps. Experiments on benchmark real datasets demonstrate that our approach can greatly improve the capability of creating realistic low-likelihood minority instances over the existing techniques without the reliance on costly additional elements. Code is available at \url{https://github.com/soobin-um/sg-minority}.
Authors: Lei Yao, Yi Wang, Moyun Liu, Lap-Pui Chau
Abstract: In recent years, transformer-based models have exhibited considerable potential in point cloud instance segmentation. Despite the promising performance achieved by existing methods, they encounter challenges such as instance query initialization problems and excessive reliance on stacked layers, rendering them incompatible with large-scale 3D scenes. This paper introduces a novel method, named SGIFormer, for 3D instance segmentation, which is composed of the Semantic-guided Mix Query (SMQ) initialization and the Geometric-enhanced Interleaving Transformer (GIT) decoder. Specifically, the principle of our SMQ initialization scheme is to leverage the predicted voxel-wise semantic information to implicitly generate the scene-aware query, yielding adequate scene prior and compensating for the learnable query set. Subsequently, we feed the formed overall query into our GIT decoder to alternately refine instance query and global scene features for further capturing fine-grained information and reducing complex design intricacies simultaneously. To emphasize geometric property, we consider bias estimation as an auxiliary task and progressively integrate shifted point coordinates embedding to reinforce instance localization. SGIFormer attains state-of-the-art performance on ScanNet V2, ScanNet200 datasets, and the challenging high-fidelity ScanNet++ benchmark, striking a balance between accuracy and efficiency. The code, weights, and demo videos are publicly available at https://rayyoh.github.io/sgiformer.
Authors: Hannes Mareen, Dimitrios Karageorgiou, Glenn Van Wallendael, Peter Lambert, Symeon Papadopoulos
Abstract: Digital image manipulation has become increasingly accessible and realistic with the advent of generative AI technologies. Recent developments allow for text-guided inpainting, making sophisticated image edits possible with minimal effort. This poses new challenges for digital media forensics. For example, diffusion model-based approaches could either splice the inpainted region into the original image, or regenerate the entire image. In the latter case, traditional image forgery localization (IFL) methods typically fail. This paper introduces the Text-Guided Inpainting Forgery (TGIF) dataset, a comprehensive collection of images designed to support the training and evaluation of image forgery localization and synthetic image detection (SID) methods. The TGIF dataset includes approximately 80k forged images, originating from popular open-source and commercial methods; SD2, SDXL, and Adobe Firefly. Using this data, we benchmark several state-of-the-art IFL and SID methods. Whereas traditional IFL methods can detect spliced images, they fail to detect regenerated inpainted images. Moreover, traditional SID may detect the regenerated inpainted images to be fake, but cannot localize the inpainted area. Finally, both types of methods fail when exposed to stronger compression, while they are less robust to modern compression algorithms, such as WEBP. As such, this work demonstrates the inefficiency of state-of-the-art detectors on local manipulations performed by modern generative approaches, and aspires to help with the development of more capable IFL and SID methods. The dataset can be downloaded at https://github.com/IDLabMedia/tgif-dataset.
Authors: Yanbo Wang, Wentao Zhao, Chuan Cao, Tianchen Deng, Jingchuan Wang, Weidong Chen
Abstract: Although LiDAR semantic segmentation advances rapidly, state-of-the-art methods often incorporate specifically designed inductive bias derived from benchmarks originating from mechanical spinning LiDAR. This can limit model generalizability to other kinds of LiDAR technologies and make hyperparameter tuning more complex. To tackle these issues, we propose a generalized framework to accommodate various types of LiDAR prevalent in the market by replacing window-attention with our sparse focal point modulation. Our SFPNet is capable of extracting multi-level contexts and dynamically aggregating them using a gate mechanism. By implementing a channel-wise information query, features that incorporate both local and global contexts are encoded. We also introduce a novel large-scale hybrid-solid LiDAR semantic segmentation dataset for robotic applications. SFPNet demonstrates competitive performance on conventional benchmarks derived from mechanical spinning LiDAR, while achieving state-of-the-art results on benchmark derived from solid-state LiDAR. Additionally, it outperforms existing methods on our novel dataset sourced from hybrid-solid LiDAR. Code and dataset are available at https://github.com/Cavendish518/SFPNet and https://www.semanticindustry.top.
URLs: https://github.com/Cavendish518/SFPNet, https://www.semanticindustry.top.
Authors: Naif Alkhunaizi, Faris Almalik, Rouqaiah Al-Refai, Muzammal Naseer, Karthik Nandakumar
Abstract: With the advent of large pre-trained transformer models, fine-tuning these models for various downstream tasks is a critical problem. Paucity of training data, the existence of data silos, and stringent privacy constraints exacerbate this fine-tuning problem in the medical imaging domain, creating a strong need for algorithms that enable collaborative fine-tuning of pre-trained models. Moreover, the large size of these models necessitates the use of parameter-efficient fine-tuning (PEFT) to reduce the communication burden in federated learning. In this work, we systematically investigate various federated PEFT strategies for adapting a Vision Transformer (ViT) model (pre-trained on a large natural image dataset) for medical image classification. Apart from evaluating known PEFT techniques, we introduce new federated variants of PEFT algorithms such as visual prompt tuning (VPT), low-rank decomposition of visual prompts, stochastic block attention fine-tuning, and hybrid PEFT methods like low-rank adaptation (LoRA)+VPT. Moreover, we perform a thorough empirical analysis to identify the optimal PEFT method for the federated setting and understand the impact of data distribution on federated PEFT, especially for out-of-domain (OOD) and non-IID data. The key insight of this study is that while most federated PEFT methods work well for in-domain transfer, there is a substantial accuracy vs. efficiency trade-off when dealing with OOD and non-IID scenarios, which is commonly the case in medical imaging. Specifically, every order of magnitude reduction in fine-tuned/exchanged parameters can lead to a 4% drop in accuracy. Thus, the initial model choice is crucial for federated PEFT. It is preferable to use medical foundation models learned from in-domain medical image data (if available) rather than general vision models.
Authors: Zeyu Wang, Zecheng Hao, Jingyu Lin, Yuchao Feng, Yufei Guo
Abstract: This study introduces a novel Remote Sensing (RS) Urban Prediction (UP) task focused on future urban planning, which aims to forecast urban layouts by utilizing information from existing urban layouts and planned change maps. To address the proposed RS UP task, we propose UP-Diff, which leverages a Latent Diffusion Model (LDM) to capture positionaware embeddings of pre-change urban layouts and planned change maps. In specific, the trainable cross-attention layers within UP-Diff's iterative diffusion modules enable the model to dynamically highlight crucial regions for targeted modifications. By utilizing our UP-Diff, designers can effectively refine and adjust future urban city plans by making modifications to the change maps in a dynamic and adaptive manner. Compared with conventional RS Change Detection (CD) methods, the proposed UP-Diff for the RS UP task avoids the requirement of paired prechange and post-change images, which enhances the practical usage in city development. Experimental results on LEVIRCD and SYSU-CD datasets show UP-Diff's ability to accurately predict future urban layouts with high fidelity, demonstrating its potential for urban planning. Code and model weights will be available upon the acceptance of the work.
Authors: Shilong Tian, Hong Chen, Chengtao Lv, Yu Liu, Jinyang Guo, Xianglong Liu, Shengxi Li, Hao Yang, Tao Xie
Abstract: Recently, video diffusion models (VDMs) have garnered significant attention due to their notable advancements in generating coherent and realistic video content. However, processing multiple frame features concurrently, coupled with the considerable model size, results in high latency and extensive memory consumption, hindering their broader application. Post-training quantization (PTQ) is an effective technique to reduce memory footprint and improve computational efficiency. Unlike image diffusion, we observe that the temporal features, which are integrated into all frame features, exhibit pronounced skewness. Furthermore, we investigate significant inter-channel disparities and asymmetries in the activation of video diffusion models, resulting in low coverage of quantization levels by individual channels and increasing the challenge of quantization. To address these issues, we introduce the first PTQ strategy tailored for video diffusion models, dubbed QVD. Specifically, we propose the High Temporal Discriminability Quantization (HTDQ) method, designed for temporal features, which retains the high discriminability of quantized features, providing precise temporal guidance for all video frames. In addition, we present the Scattered Channel Range Integration (SCRI) method which aims to improve the coverage of quantization levels across individual channels. Experimental validations across various models, datasets, and bit-width settings demonstrate the effectiveness of our QVD in terms of diverse metrics. In particular, we achieve near-lossless performance degradation on W8A8, outperforming the current methods by 205.12 in FVD.
Authors: Xiaotong Lin, Tianming Liang, Jianhuang Lai, Jian-Fang Hu
Abstract: Human trajectory prediction is a practical task of predicting the future positions of pedestrians on the road, which typically covers all temporal ranges from short-term to long-term within a trajectory. However, existing works attempt to address the entire trajectory prediction with a singular, uniform training paradigm, neglecting the distinction between short-term and long-term dynamics in human trajectories. To overcome this limitation, we introduce a novel Progressive Pretext Task learning (PPT) framework, which progressively enhances the model's capacity of capturing short-term dynamics and long-term dependencies for the final entire trajectory prediction. Specifically, we elaborately design three stages of training tasks in the PPT framework. In the first stage, the model learns to comprehend the short-term dynamics through a stepwise next-position prediction task. In the second stage, the model is further enhanced to understand long-term dependencies through a destination prediction task. In the final stage, the model aims to address the entire future trajectory task by taking full advantage of the knowledge from previous stages. To alleviate the knowledge forgetting, we further apply a cross-task knowledge distillation. Additionally, we design a Transformer-based trajectory predictor, which is able to achieve highly efficient two-step reasoning by integrating a destination-driven prediction strategy and a group of learnable prompt embeddings. Extensive experiments on popular benchmarks have demonstrated that our proposed approach achieves state-of-the-art performance with high efficiency. Code is available at https://github.com/iSEE-Laboratory/PPT.
Authors: Guillermo Jimenez-Perez, Pedro Osorio, Josef Cersovsky, Javier Montalt-Tordera, Jens Hooge, Steffen Vogler, Sadegh Mohammadi
Abstract: Diffusion models (DMs) have emerged as powerful foundation models for a variety of tasks, with a large focus in synthetic image generation. However, their requirement of large annotated datasets for training limits their applicability in medical imaging, where datasets are typically smaller and sparsely annotated. We introduce DiNO-Diffusion, a self-supervised method for training latent diffusion models (LDMs) that conditions the generation process on image embeddings extracted from DiNO. By eliminating the reliance on annotations, our training leverages over 868k unlabelled images from public chest X-Ray (CXR) datasets. Despite being self-supervised, DiNO-Diffusion shows comprehensive manifold coverage, with FID scores as low as 4.7, and emerging properties when evaluated in downstream tasks. It can be used to generate semantically-diverse synthetic datasets even from small data pools, demonstrating up to 20% AUC increase in classification performance when used for data augmentation. Images were generated with different sampling strategies over the DiNO embedding manifold and using real images as a starting point. Results suggest, DiNO-Diffusion could facilitate the creation of large datasets for flexible training of downstream AI models from limited amount of real data, while also holding potential for privacy preservation. Additionally, DiNO-Diffusion demonstrates zero-shot segmentation performance of up to 84.4% Dice score when evaluating lung lobe segmentation. This evidences good CXR image-anatomy alignment, akin to segmenting using textual descriptors on vanilla DMs. Finally, DiNO-Diffusion can be easily adapted to other medical imaging modalities or state-of-the-art diffusion models, opening the door for large-scale, multi-domain image generation pipelines for medical imaging.
Authors: Weimin Wang, Yingxu Deng, Zezeng Li, Yu Liu, Na Lei
Abstract: This paper introduces a novel method for reconstructing meshes from sparse point clouds by predicting edge connection. Existing implicit methods usually produce superior smooth and watertight meshes due to the isosurface extraction algorithms~(e.g., Marching Cubes). However, these methods become memory and computationally intensive with increasing resolution. Explicit methods are more efficient by directly forming the face from points. Nevertheless, the challenge of selecting appropriate faces from enormous candidates often leads to undesirable faces and holes. Moreover, the reconstruction performance of both approaches tends to degrade when the point cloud gets sparse. To this end, we propose MEsh Reconstruction via edGE~(MergeNet), which converts mesh reconstruction into local connectivity prediction problems. Specifically, MergeNet learns to extract the features of candidate edges and regress their distances to the underlying surface. Consequently, the predicted distance is utilized to filter out edges that lay on surfaces. Finally, the meshes are reconstructed by refining the triangulations formed by these edges. Extensive experiments on synthetic and real-scanned datasets demonstrate the superiority of MergeNet to SoTA explicit methods.
Authors: Daniel Braun, Remco Chang, Michael Gleicher, Tatiana von Landesberger
Abstract: Visual validation of regression models in scatterplots is a common practice for assessing model quality, yet its efficacy remains unquantified. We conducted two empirical experiments to investigate individuals' ability to visually validate linear regression models (linear trends) and to examine the impact of common visualization designs on validation quality. The first experiment showed that the level of accuracy for visual estimation of slope (i.e., fitting a line to data) is higher than for visual validation of slope (i.e., accepting a shown line). Notably, we found bias toward slopes that are "too steep" in both cases. This lead to novel insights that participants naturally assessed regression with orthogonal distances between the points and the line (i.e., ODR regression) rather than the common vertical distances (OLS regression). In the second experiment, we investigated whether incorporating common designs for regression visualization (error lines, bounding boxes, and confidence intervals) would improve visual validation. Even though error lines reduced validation bias, results failed to show the desired improvements in accuracy for any design. Overall, our findings suggest caution in using visual model validation for linear trends in scatterplots.
Authors: Zhengcong Fei, Mingyuan Fan, Changqian Yu, Debang Li, Junshi Huang
Abstract: In this paper, we present DiT-MoE, a sparse version of the diffusion Transformer, that is scalable and competitive with dense networks while exhibiting highly optimized inference. The DiT-MoE includes two simple designs: shared expert routing and expert-level balance loss, thereby capturing common knowledge and reducing redundancy among the different routed experts. When applied to conditional image generation, a deep analysis of experts specialization gains some interesting observations: (i) Expert selection shows preference with spatial position and denoising time step, while insensitive with different class-conditional information; (ii) As the MoE layers go deeper, the selection of experts gradually shifts from specific spacial position to dispersion and balance. (iii) Expert specialization tends to be more concentrated at the early time step and then gradually uniform after half. We attribute it to the diffusion process that first models the low-frequency spatial information and then high-frequency complex information. Based on the above guidance, a series of DiT-MoE experimentally achieves performance on par with dense networks yet requires much less computational load during inference. More encouragingly, we demonstrate the potential of DiT-MoE with synthesized image data, scaling diffusion model at a 16.5B parameter that attains a new SoTA FID-50K score of 1.80 in 512$\times$512 resolution settings. The project page: https://github.com/feizc/DiT-MoE.
Authors: Han Nie, Bin Luo, Jun Liu, Zhitao Fu, Weixing Liu, Xin Su
Abstract: We present REMM, a rotation-equivariant framework for end-to-end multimodal image matching, which fully encodes rotational differences of descriptors in the whole matching pipeline. Previous learning-based methods mainly focus on extracting modal-invariant descriptors, while consistently ignoring the rotational invariance. In this paper, we demonstrate that our REMM is very useful for multimodal image matching, including multimodal feature learning module and cyclic shift module. We first learn modal-invariant features through the multimodal feature learning module. Then, we design the cyclic shift module to rotationally encode the descriptors, greatly improving the performance of rotation-equivariant matching, which makes them robust to any angle. To validate our method, we establish a comprehensive rotation and scale-matching benchmark for evaluating the anti-rotation performance of multimodal images, which contains a combination of multi-angle and multi-scale transformations from four publicly available datasets. Extensive experiments show that our method outperforms existing methods in benchmarking and generalizes well to independent datasets. Additionally, we conducted an in-depth analysis of the key components of the REMM to validate the improvements brought about by the cyclic shift module. Code and dataset at https://github.com/HanNieWHU/REMM.
Authors: Guoliang You, Xiaomeng Chu, Yifan Duan, Wenyu Zhang, Xingchen Li, Sha Zhang, Yao Li, Jianmin Ji, Yanyong Zhang
Abstract: When planning for autonomous driving, it is crucial to consider essential traffic elements such as lanes, intersections, traffic regulations, and dynamic agents. However, they are often overlooked by the traditional end-to-end planning methods, likely leading to inefficiencies and non-compliance with traffic regulations. In this work, we endeavor to integrate the perception of these elements into the planning task. To this end, we propose Perception Helps Planning (PHP), a novel framework that reconciles lane-level planning with perception. This integration ensures that planning is inherently aligned with traffic constraints, thus facilitating safe and efficient driving. Specifically, PHP focuses on both edges of a lane for planning and perception purposes, taking into consideration the 3D positions of both lane edges and attributes for lane intersections, lane directions, lane occupancy, and planning. In the algorithmic design, the process begins with the transformer encoding multi-camera images to extract the above features and predicting lane-level perception results. Next, the hierarchical feature early fusion module refines the features for predicting planning attributes. Finally, the double-edge interpreter utilizes a late-fusion process specifically designed to integrate lane-level perception and planning information, culminating in the generation of vehicle control signals. Experiments on three Carla benchmarks show significant improvements in driving score of 27.20%, 33.47%, and 15.54% over existing algorithms, respectively, achieving the state-of-the-art performance, with the system operating up to 22.57 FPS.
Authors: Marcella Astrid, Enjie Ghorbel, Djamila Aouada
Abstract: In this paper, we propose an enhanced audio-visual deep detection method. Recent methods in audio-visual deepfake detection mostly assess the synchronization between audio and visual features. Although they have shown promising results, they are based on the maximization/minimization of isolated feature distances without considering feature statistics. Moreover, they rely on cumbersome deep learning architectures and are heavily dependent on empirically fixed hyperparameters. Herein, to overcome these limitations, we propose: (1) a statistical feature loss to enhance the discrimination capability of the model, instead of relying solely on feature distances; (2) using the waveform for describing the audio as a replacement of frequency-based representations; (3) a post-processing normalization of the fakeness score; (4) the use of shallower network for reducing the computational complexity. Experiments on the DFDC and FakeAVCeleb datasets demonstrate the relevance of the proposed method.
Authors: Sunny Gupta, Amit Sethi
Abstract: Federated Learning (FL) offers a privacy-preserving approach to train models on decentralized data. Its potential in healthcare is significant, but challenges arise due to cross-client variations in medical image data, exacerbated by limited annotations. This paper introduces Cross-Client Variations Adaptive Federated Learning (CCVA-FL) to address these issues. CCVA-FL aims to minimize cross-client variations by transforming images into a common feature space. It involves expert annotation of a subset of images from each client, followed by the selection of a client with the least data complexity as the target. Synthetic medical images are then generated using Scalable Diffusion Models with Transformers (DiT) based on the target client's annotated images. These synthetic images, capturing diversity and representing the original data, are shared with other clients. Each client then translates its local images into the target image space using image-to-image translation. The translated images are subsequently used in a federated learning setting to develop a server model. Our results demonstrate that CCVA-FL outperforms Vanilla Federated Averaging by effectively addressing data distribution differences across clients without compromising privacy.
Authors: Xiaodong Li, Wenchao Du, Hongyu Yang
Abstract: In this paper, we present our solution and experiment result for the Multi-Task Learning Challenge of the 7th Affective Behavior Analysis in-the-wild(ABAW7) Competition. This challenge consists of three tasks: action unit detection, facial expression recognition, and valance-arousal estimation. We address the research problems of this challenge from three aspects: 1)For learning robust visual feature representations, we introduce the pre-trained large model Dinov2. 2) To adaptively extract the required features of eack task, we design a task-adaptive block that performs cross-attention between a set of learnable query vectors and pre-extracted features. 3) By proposing the AU-assisted Graph Convolutional Network(AU-GCN), we make full use of the correlation information between AUs to assist in solving the EXPR and VA tasks. Finally, we achieve the evaluation measure of \textbf{1.2542} on the validation set provided by the organizers.
Authors: Dominik Winter, Nicolas Triltsch, Marco Rosati, Anatoliy Shumilov, Ziya Kokaragac, Yuri Popov, Thomas Padel, Laura Sebastian Monasor, Ross Hill, Markus Schick, Nicolas Brieu
Abstract: Creating in-silico data with generative AI promises a cost-effective alternative to staining, imaging, and annotating whole slide images in computational pathology. Diffusion models are the state-of-the-art solution for generating in-silico images, offering unparalleled fidelity and realism. Using appearance transfer diffusion models allows for zero-shot image generation, facilitating fast application and making model training unnecessary. However current appearance transfer diffusion models are designed for natural images, where the main task is to transfer the foreground object from an origin to a target domain, while the background is of insignificant importance. In computational pathology, specifically in oncology, it is however not straightforward to define which objects in an image should be classified as foreground and background, as all objects in an image may be of critical importance for the detailed understanding the tumor micro-environment. We contribute to the applicability of appearance transfer diffusion models to immunohistochemistry-stained images by modifying the appearance transfer guidance to alternate between class-specific AdaIN feature statistics matchings using existing segmentation masks. The performance of the proposed method is demonstrated on the downstream task of supervised epithelium segmentation, showing that the number of manual annotations required for model training can be reduced by 75%, outperforming the baseline approach. Additionally, we consulted with a certified pathologist to investigate future improvements. We anticipate this work to inspire the application of zero-shot diffusion models in computational pathology, providing an efficient method to generate in-silico images with unmatched fidelity and realism, which prove meaningful for downstream tasks, such as training existing deep learning models or finetuning foundation models.
Authors: Shinjeong Kim, Marc Pollefeys, Daniel Barath
Abstract: This work addresses the challenge of sub-pixel accuracy in detecting 2D local features, a cornerstone problem in computer vision. Despite the advancements brought by neural network-based methods like SuperPoint and ALIKED, these modern approaches lag behind classical ones such as SIFT in keypoint localization accuracy due to their lack of sub-pixel precision. We propose a novel network that enhances any detector with sub-pixel precision by learning an offset vector for detected features, thereby eliminating the need for designing specialized sub-pixel accurate detectors. This optimization directly minimizes test-time evaluation metrics like relative pose error. Through extensive testing with both nearest neighbors matching and the recent LightGlue matcher across various real-world datasets, our method consistently outperforms existing methods in accuracy. Moreover, it adds only around 7 ms to the time of a particular detector. The code is available at https://github.com/KimSinjeong/keypt2subpx .
Authors: Shi-Xue Zhang, Hongfa Wang, Xiaobin Zhu, Weibo Gu, Tianjin Zhang, Chun Yang, Wei Liu, Xu-Cheng Yin
Abstract: Video-language alignment is a crucial multi-modal task that benefits various downstream applications, e.g., video-text retrieval and video question answering. Existing methods either utilize multi-modal information in video-text pairs or apply global and local alignment techniques to promote alignment precision. However, these methods often fail to fully explore the spatio-temporal relationships among vision tokens within video and across different video-text pairs. In this paper, we propose a novel Spatio-Temporal Graph Transformer module to uniformly learn spatial and temporal contexts for video-language alignment pre-training (dubbed STGT). Specifically, our STGT combines spatio-temporal graph structure information with attention in transformer block, effectively utilizing the spatio-temporal contexts. In this way, we can model the relationships between vision tokens, promoting video-text alignment precision for benefiting downstream tasks. In addition, we propose a self-similarity alignment loss to explore the inherent self-similarity in the video and text. With the initial optimization achieved by contrastive learning, it can further promote the alignment accuracy between video and text. Experimental results on challenging downstream tasks, including video-text retrieval and video question answering, verify the superior performance of our method.
Authors: Xiaoshuai Hao, Ruikai Li, Hui Zhang, Dingzhe Li, Rong Yin, Sangil Jung, Seung-In Park, ByungIn Yoo, Haimei Zhao, Jing Zhang
Abstract: Online high-definition (HD) map construction is an important and challenging task in autonomous driving. Recently, there has been a growing interest in cost-effective multi-view camera-based methods without relying on other sensors like LiDAR. However, these methods suffer from a lack of explicit depth information, necessitating the use of large models to achieve satisfactory performance. To address this, we employ the Knowledge Distillation (KD) idea for efficient HD map construction for the first time and introduce a novel KD-based approach called MapDistill to transfer knowledge from a high-performance camera-LiDAR fusion model to a lightweight camera-only model. Specifically, we adopt the teacher-student architecture, i.e., a camera-LiDAR fusion model as the teacher and a lightweight camera model as the student, and devise a dual BEV transform module to facilitate cross-modal knowledge distillation while maintaining cost-effective camera-only deployment. Additionally, we present a comprehensive distillation scheme encompassing cross-modal relation distillation, dual-level feature distillation, and map head distillation. This approach alleviates knowledge transfer challenges between modalities, enabling the student model to learn improved feature representations for HD map construction. Experimental results on the challenging nuScenes dataset demonstrate the effectiveness of MapDistill, surpassing existing competitors by over 7.7 mAP or 4.5X speedup.
Authors: Yunbin Tu, Liang Li, Li Su, Chenggang Yan, Qingming Huang
Abstract: Change captioning aims to succinctly describe the semantic change between a pair of similar images, while being immune to distractors (illumination and viewpoint changes). Under these distractors, unchanged objects often appear pseudo changes about location and scale, and certain objects might overlap others, resulting in perturbational and discrimination-degraded features between two images. However, most existing methods directly capture the difference between them, which risk obtaining error-prone difference features. In this paper, we propose a distractors-immune representation learning network that correlates the corresponding channels of two image representations and decorrelates different ones in a self-supervised manner, thus attaining a pair of stable image representations under distractors. Then, the model can better interact them to capture the reliable difference features for caption generation. To yield words based on the most related difference features, we further design a cross-modal contrastive regularization, which regularizes the cross-modal alignment by maximizing the contrastive alignment between the attended difference features and generated words. Extensive experiments show that our method outperforms the state-of-the-art methods on four public datasets. The code is available at https://github.com/tuyunbin/DIRL.
Authors: Haodong Duan, Junming Yang, Yuxuan Qiao, Xinyu Fang, Lin Chen, Yuan Liu, Xiaoyi Dong, Yuhang Zang, Pan Zhang, Jiaqi Wang, Dahua Lin, Kai Chen
Abstract: We present VLMEvalKit: an open-source toolkit for evaluating large multi-modality models based on PyTorch. The toolkit aims to provide a user-friendly and comprehensive framework for researchers and developers to evaluate existing multi-modality models and publish reproducible evaluation results. In VLMEvalKit, we implement over 70 different large multi-modality models, including both proprietary APIs and open-source models, as well as more than 20 different multi-modal benchmarks. By implementing a single interface, new models can be easily added to the toolkit, while the toolkit automatically handles the remaining workloads, including data preparation, distributed inference, prediction post-processing, and metric calculation. Although the toolkit is currently mainly used for evaluating large vision-language models, its design is compatible with future updates that incorporate additional modalities, such as audio and video. Based on the evaluation results obtained with the toolkit, we host OpenVLM Leaderboard, a comprehensive leaderboard to track the progress of multi-modality learning research. The toolkit is released at https://github.com/open-compass/VLMEvalKit and is actively maintained.
Authors: Xiuquan Hou, Meiqin Liu, Senlin Zhang, Ping Wei, Badong Chen, Xuguang Lan
Abstract: This paper presents a general scheme for enhancing the convergence and performance of DETR (DEtection TRansformer). We investigate the slow convergence problem in transformers from a new perspective, suggesting that it arises from the self-attention that introduces no structural bias over inputs. To address this issue, we explore incorporating position relation prior as attention bias to augment object detection, following the verification of its statistical significance using a proposed quantitative macroscopic correlation (MC) metric. Our approach, termed Relation-DETR, introduces an encoder to construct position relation embeddings for progressive attention refinement, which further extends the traditional streaming pipeline of DETR into a contrastive relation pipeline to address the conflicts between non-duplicate predictions and positive supervision. Extensive experiments on both generic and task-specific datasets demonstrate the effectiveness of our approach. Under the same configurations, Relation-DETR achieves a significant improvement (+2.0% AP compared to DINO), state-of-the-art performance (51.7% AP for 1x and 52.1% AP for 2x settings), and a remarkably faster convergence speed (over 40% AP with only 2 training epochs) than existing DETR detectors on COCO val2017. Moreover, the proposed relation encoder serves as a universal plug-in-and-play component, bringing clear improvements for theoretically any DETR-like methods. Furthermore, we introduce a class-agnostic detection dataset, SA-Det-100k. The experimental results on the dataset illustrate that the proposed explicit position relation achieves a clear improvement of 1.3% AP, highlighting its potential towards universal object detection. The code and dataset are available at https://github.com/xiuqhou/Relation-DETR.
Authors: Jinming Liu, Ruoyu Feng, Yunpeng Qi, Qiuyu Chen, Zhibo Chen, Wenjun Zeng, Xin Jin
Abstract: Recently, the field of Image Coding for Machines (ICM) has garnered heightened interest and significant advances thanks to the rapid progress of learning-based techniques for image compression and analysis. Previous studies often require training separate codecs to support various bitrate levels, machine tasks, and networks, thus lacking both flexibility and practicality. To address these challenges, we propose a rate-distortion-cognition controllable versatile image compression, which method allows the users to adjust the bitrate (i.e., Rate), image reconstruction quality (i.e., Distortion), and machine task accuracy (i.e., Cognition) with a single neural model, achieving ultra-controllability. Specifically, we first introduce a cognition-oriented loss in the primary compression branch to train a codec for diverse machine tasks. This branch attains variable bitrate by regulating quantization degree through the latent code channels. To further enhance the quality of the reconstructed images, we employ an auxiliary branch to supplement residual information with a scalable bitstream. Ultimately, two branches use a `$\beta x + (1 - \beta) y$' interpolation strategy to achieve a balanced cognition-distortion trade-off. Extensive experiments demonstrate that our method yields satisfactory ICM performance and flexible Rate-Distortion-Cognition controlling.
Authors: Shumei Liu, Haiting Huang, Mathias Zinnen, Andreas Maier, Vincent Christlein
Abstract: Olfaction, often overlooked in cultural heritage studies, holds profound significance in shaping human experiences and identities. Examining historical depictions of olfactory scenes can offer valuable insights into the role of smells in history. We show that a transfer-learning approach using weakly labeled training data can remarkably improve the classification of fragrant spaces and, more generally, artistic scene depictions. We fine-tune Places365-pre-trained models by querying two cultural heritage data sources and using the search terms as supervision signal. The models are evaluated on two manually corrected test splits. This work lays a foundation for further exploration of fragrant spaces recognition and artistic scene classification. All images and labels are released as the ArtPlaces dataset at https://zenodo.org/doi/10.5281/zenodo.11584328.
Authors: Suhwan Cho, Minhyeok Lee, Jungho Lee, Donghyeong Kim, Seunghoon Lee, Sungmin Woo, Sangyoun Lee
Abstract: Unsupervised video object segmentation (VOS), also known as video salient object detection, aims to detect the most prominent object in a video at the pixel level. Recently, two-stream approaches that leverage both RGB images and optical flow maps have gained significant attention. However, the limited amount of training data remains a substantial challenge. In this study, we propose a novel data generation method that simulates fake optical flows from single images, thereby creating large-scale training data for stable network learning. Inspired by the observation that optical flow maps are highly dependent on depth maps, we generate fake optical flows by refining and augmenting the estimated depth maps of each image. By incorporating our simulated image-flow pairs, we achieve new state-of-the-art performance on all public benchmark datasets without relying on complex modules. We believe that our data generation method represents a potential breakthrough for future VOS research.
Authors: Chen Ju, Haicheng Wang, Haozhe Cheng, Xu Chen, Zhonghua Zhai, Weilin Huang, Jinsong Lan, Shuai Xiao, Bo Zheng
Abstract: Vision-Language Large Models (VLMs) recently become primary backbone of AI, due to the impressive performance. However, their expensive computation costs, i.e., throughput and delay, impede potentials in the real-world scenarios. To achieve acceleration for VLMs, most existing methods focus on the model perspective: pruning, distillation, quantization, but completely overlook the data-perspective redundancy. To fill the overlook, this paper pioneers the severity of data redundancy, and designs one plug-and-play Turbo module guided by information degree to prune inefficient tokens from visual or textual data. In pursuit of efficiency-performance trade-offs, information degree takes two crucial factors into consideration: mutual redundancy and semantic value. Concretely, the former evaluates data duplication between sequential tokens; while the latter evaluates each token by its contribution to the overall semantics. As a result, tokens with high information degree carry less redundancy and stronger semantics. For VLMs' calculation, Turbo works as a user-friendly plug-in that sorts data referring to information degree, utilizing only top-level ones to save costs. Its advantages are multifaceted, e.g., being generally compatible to various VLMs across understanding and generation, simple use without re-training and trivial engineering efforts. On multiple VLMs benchmarks, we fully experiment to demonstrate the good acceleration of Turbo, under negligible performance drop.
Authors: Hongxiao Yu, Yuqi Wang, Yuntao Chen, Zhaoxiang Zhang
Abstract: Camera-based 3D occupancy prediction has recently garnered increasing attention in outdoor driving scenes. However, research in indoor scenes remains relatively unexplored. The core differences in indoor scenes lie in the complexity of scene scale and the variance in object size. In this paper, we propose a novel method, named ISO, for predicting indoor scene occupancy using monocular images. ISO harnesses the advantages of a pretrained depth model to achieve accurate depth predictions. Furthermore, we introduce the Dual Feature Line of Sight Projection (D-FLoSP) module within ISO, which enhances the learning of 3D voxel features. To foster further research in this domain, we introduce Occ-ScanNet, a large-scale occupancy benchmark for indoor scenes. With a dataset size 40 times larger than the NYUv2 dataset, it facilitates future scalable research in indoor scene analysis. Experimental results on both NYUv2 and Occ-ScanNet demonstrate that our method achieves state-of-the-art performance. The dataset and code are made publicly at https://github.com/hongxiaoy/ISO.git.
Authors: Josh Veitch-Michaelis, Andrew Cottam, Daniella Schweizer, Eben N. Broadbent, David Dao, Ce Zhang, Angelica Almeyda Zambrano, Simeon Max
Abstract: Accurately quantifying tree cover is an important metric for ecosystem monitoring and for assessing progress in restored sites. Recent works have shown that deep learning-based segmentation algorithms are capable of accurately mapping trees at country and continental scales using high-resolution aerial and satellite imagery. Mapping at high (ideally sub-meter) resolution is necessary to identify individual trees, however there are few open-access datasets containing instance level annotations and those that exist are small or not geographically diverse. We present a novel open-access dataset for individual tree crown delineation (TCD) in high-resolution aerial imagery sourced from OpenAerialMap (OAM). Our dataset, OAM-TCD, comprises 5072 2048x2048 px images at 10 cm/px resolution with associated human-labeled instance masks for over 280k individual and 56k groups of trees. By sampling imagery from around the world, we are able to better capture the diversity and morphology of trees in different terrestrial biomes and in both urban and natural environments. Using our dataset, we train reference instance and semantic segmentation models that compare favorably to existing state-of-the-art models. We assess performance through k-fold cross-validation and comparison with existing datasets; additionally we demonstrate compelling results on independent aerial imagery captured over Switzerland and compare to municipal tree inventories and LIDAR-derived canopy maps in the city of Zurich. Our dataset, models and training/benchmark code are publicly released under permissive open-source licenses: Creative Commons (majority CC BY 4.0), and Apache 2.0 respectively.
Authors: Chen Zhao, Weiling Cai, ChengWei Hu, Zheng Yuan
Abstract: To tackle the difficulties in fitting paired real-world data for single image deraining (SID), recent unsupervised methods have achieved notable success. However, these methods often struggle to generate high-quality, rain-free images due to a lack of attention to semantic representation and image content, resulting in ineffective separation of content from the rain layer. In this paper, we propose a novel cycle contrastive generative adversarial network for unsupervised SID, called CCLGAN. This framework combines cycle contrastive learning (CCL) and location contrastive learning (LCL). CCL improves image reconstruction and rain-layer removal by bringing similar features closer and pushing dissimilar features apart in both semantic and discriminative spaces. At the same time, LCL preserves content information by constraining mutual information at the same location across different exemplars. CCLGAN shows superior performance, as extensive experiments demonstrate the benefits of CCLGAN and the effectiveness of its components.
Authors: Pandiyaraju V, Shravan Venkatraman, Abeshek A, Pavan Kumar S, Aravintakshan S A, Senthil Kumar A M, Kannan A
Abstract: Farmers face various challenges when it comes to identifying diseases in rice leaves during their early stages of growth, which is a major reason for poor produce. Therefore, early and accurate disease identification is important in agriculture to avoid crop loss and improve cultivation. In this research, we propose a novel hybrid deep learning (DL) classifier designed by extending the Squeeze-and-Excitation network architecture with a channel attention mechanism and the Swish ReLU activation function. The channel attention mechanism in our proposed model identifies the most important feature channels required for classification during feature extraction and selection. The dying ReLU problem is mitigated by utilizing the Swish ReLU activation function, and the Squeeze-andExcitation blocks improve information propagation and cross-channel interaction. Upon evaluation, our model achieved a high F1-score of 99.76% and an accuracy of 99.74%, surpassing the performance of existing models. These outcomes demonstrate the potential of state-of-the-art DL techniques in agriculture, contributing to the advancement of more efficient and reliable disease detection systems.
Authors: Truong Thanh Hung Nguyen, Phuc Truong Loc Nguyen, Hung Cao
Abstract: Recent advancements in deep learning have significantly improved visual quality inspection and predictive maintenance within industrial settings. However, deploying these technologies on low-resource edge devices poses substantial challenges due to their high computational demands and the inherent complexity of Explainable AI (XAI) methods. This paper addresses these challenges by introducing a novel XAI-integrated Visual Quality Inspection framework that optimizes the deployment of semantic segmentation models on low-resource edge devices. Our framework incorporates XAI and the Large Vision Language Model to deliver human-centered interpretability through visual and textual explanations to end-users. This is crucial for end-user trust and model interpretability. We outline a comprehensive methodology consisting of six fundamental modules: base model fine-tuning, XAI-based explanation generation, evaluation of XAI approaches, XAI-guided data augmentation, development of an edge-compatible model, and the generation of understandable visual and textual explanations. Through XAI-guided data augmentation, the enhanced model incorporating domain expert knowledge with visual and textual explanations is successfully deployed on mobile devices to support end-users in real-world scenarios. Experimental results showcase the effectiveness of the proposed framework, with the mobile model achieving competitive accuracy while significantly reducing model size. This approach paves the way for the broader adoption of reliable and interpretable AI tools in critical industrial applications, where decisions must be both rapid and justifiable.
Authors: Shuang Li, Yibing Wang, Jian Gao, Chulhong Kim, Seongwook Choi, Yu Zhang, Qian Chen, Yao Yao, Changhui Li
Abstract: High-quality 3D photoacoustic imaging (PAI) reconstruction under sparse view or limited view has long been challenging. Traditional 3D iterative-based reconstruction methods suffer from both slow speed and high memory consumption. Recently, in computer graphics, the differentiable rendering has made significant progress, particularly with the rise of 3D Gaussian Splatting. Inspired by these, we introduce differentiable radiation into PAI, developing a novel reconstruction algorithm: the Sliding Ball Adaptive Growth algorithm (SlingBAG) for 3D PAI, which shows ability in high-quality 3D PAI reconstruction both under extremely sparse view and limited view. We established the point cloud dataset in PAI, and used unique differentiable rapid radiator based on the spherical decomposition strategy and the randomly initialized point cloud adaptively optimized according to sparse sensor data. Each point undergoes updates in 3D coordinates, initial pressure, and resolution (denoted by the radius of ball). Points undergo adaptive growth during iterative process, including point destroying, splitting and duplicating along the gradient of their positions, manifesting the sliding ball effect. Finally, our point cloud to voxel grid shader renders the final reconstruction results. Simulation and in vivo experiments demonstrate that our SlingBAG reconstruction result's SNR can be more than 40 dB under extremely sparse view, while the SNR of traditional back-projection algorithm's result is less than 20 dB. Moreover, the result of SlingBAG's structural similarity to the ground truth is significantly higher, with an SSIM value of 95.6%. Notably, our differentiable rapid radiator can conduct forward PA simulation in homogeneous, non-viscous media substantially faster than current methods that numerically simulate the wave propagation, such as k-Wave. The dataset and all code will be open source.
Authors: Seokhun Choi, Hyeonseop Song, Jaechul Kim, Taehyeong Kim, Hoseok Do
Abstract: Interactive segmentation of 3D Gaussians opens a great opportunity for real-time manipulation of 3D scenes thanks to the real-time rendering capability of 3D Gaussian Splatting. However, the current methods suffer from time-consuming post-processing to deal with noisy segmentation output. Also, they struggle to provide detailed segmentation, which is important for fine-grained manipulation of 3D scenes. In this study, we propose Click-Gaussian, which learns distinguishable feature fields of two-level granularity, facilitating segmentation without time-consuming post-processing. We delve into challenges stemming from inconsistently learned feature fields resulting from 2D segmentation obtained independently from a 3D scene. 3D segmentation accuracy deteriorates when 2D segmentation results across the views, primary cues for 3D segmentation, are in conflict. To overcome these issues, we propose Global Feature-guided Learning (GFL). GFL constructs the clusters of global feature candidates from noisy 2D segments across the views, which smooths out noises when training the features of 3D Gaussians. Our method runs in 10 ms per click, 15 to 130 times as fast as the previous methods, while also significantly improving segmentation accuracy. Our project page is available at https://seokhunchoi.github.io/Click-Gaussian
Authors: Nikolaos Giakoumoglou, Tania Stathaki
Abstract: Knowledge distillation (KD) involves transferring the knowledge from one neural network to another, often from a larger, well-trained model (teacher) to a smaller, more efficient model (student). Traditional KD methods minimize the Kullback-Leibler (KL) divergence between the probabilistic outputs of the teacher and student networks. However, this approach often overlooks crucial structural knowledge embedded within the teacher's network. In this paper, we introduce Invariant Consistency Distillation (ICD), a novel methodology designed to enhance KD by ensuring that the student model's representations are consistent with those of the teacher. Our approach combines contrastive learning with an explicit invariance penalty, capturing significantly more information from the teacher's representation of the data. Our results on CIFAR-100 demonstrate that ICD outperforms traditional KD techniques and surpasses 13 state-of-the-art methods. In some cases, the student model even exceeds the teacher model in terms of accuracy. Furthermore, we successfully transfer our method to other datasets, including Tiny ImageNet and STL-10. The code will be made public soon.
Authors: Vasco Ramos, Yonatan Bitton, Michal Yarom, Idan Szpektor, Joao Magalhaes
Abstract: Action-centric sequence descriptions like recipe instructions and do-it-yourself projects include non-linear patterns in which the next step may require to be visually consistent not on the immediate previous step but on earlier steps. Current video synthesis approaches fail to generate consistent multi-scene videos for such task descriptions. We propose a contrastive sequential video diffusion method that selects the most suitable previously generated scene to guide and condition the denoising process of the next scene. The result is a multi-scene video that is grounded in the scene descriptions and coherent w.r.t the scenes that require consistent visualisation. Our experiments with real-world data demonstrate the practicality and improved consistency of our model compared to prior work.
Authors: Juncheng Ma, Peiwen Sun, Yaoting Wang, Di Hu
Abstract: Audio-Visual Segmentation (AVS) aims to achieve pixel-level localization of sound sources in videos, while Audio-Visual Semantic Segmentation (AVSS), as an extension of AVS, further pursues semantic understanding of audio-visual scenes. However, since the AVSS task requires the establishment of audio-visual correspondence and semantic understanding simultaneously, we observe that previous methods have struggled to handle this mashup of objectives in end-to-end training, resulting in insufficient learning and sub-optimization. Therefore, we propose a two-stage training strategy called \textit{Stepping Stones}, which decomposes the AVSS task into two simple subtasks from localization to semantic understanding, which are fully optimized in each stage to achieve step-by-step global optimization. This training strategy has also proved its generalization and effectiveness on existing methods. To further improve the performance of AVS tasks, we propose a novel framework Adaptive Audio Visual Segmentation, in which we incorporate an adaptive audio query generator and integrate masked attention into the transformer decoder, facilitating the adaptive fusion of visual and audio features. Extensive experiments demonstrate that our methods achieve state-of-the-art results on all three AVS benchmarks. The project homepage can be accessed at https://gewu-lab.github.io/stepping_stones/.
Authors: Zhuoxiao Li, Shanliang Yao, Yijie Chu, Angel F. Garcia-Fernandez, Yong Yue, Eng Gee Lim, Xiaohui Zhu
Abstract: In the rapidly evolving field of 3D reconstruction, 3D Gaussian Splatting (3DGS) and 2D Gaussian Splatting (2DGS) represent significant advancements. Although 2DGS compresses 3D Gaussian primitives into 2D Gaussian surfels to effectively enhance mesh extraction quality, this compression can potentially lead to a decrease in rendering quality. Additionally, unreliable densification processes and the calculation of depth through the accumulation of opacity can compromise the detail of mesh extraction. To address this issue, we introduce MVG-Splatting, a solution guided by Multi-View considerations. Specifically, we integrate an optimized method for calculating normals, which, combined with image gradients, helps rectify inconsistencies in the original depth computations. Additionally, utilizing projection strategies akin to those in Multi-View Stereo (MVS), we propose an adaptive quantile-based method that dynamically determines the level of additional densification guided by depth maps, from coarse to fine detail. Experimental evidence demonstrates that our method not only resolves the issues of rendering quality degradation caused by depth discrepancies but also facilitates direct mesh extraction from dense Gaussian point clouds using the Marching Cubes algorithm. This approach significantly enhances the overall fidelity and accuracy of the 3D reconstruction process, ensuring that both the geometric details and visual quality.
Authors: Nir Barel, Ron Shapira Weber, Nir Mualem, Shahaf E. Finder, Oren Freifeld
Abstract: The unsupervised task of Joint Alignment (JA) of images is beset by challenges such as high complexity, geometric distortions, and convergence to poor local or even global optima. Although Vision Transformers (ViT) have recently provided valuable features for JA, they fall short of fully addressing these issues. Consequently, researchers frequently depend on expensive models and numerous regularization terms, resulting in long training times and challenging hyperparameter tuning. We introduce the Spatial Joint Alignment Model (SpaceJAM), a novel approach that addresses the JA task with efficiency and simplicity. SpaceJAM leverages a compact architecture with only 16K trainable parameters and uniquely operates without the need for regularization or atlas maintenance. Evaluations on SPair-71K and CUB datasets demonstrate that SpaceJAM matches the alignment capabilities of existing methods while significantly reducing computational demands and achieving at least a 10x speedup. SpaceJAM sets a new standard for rapid and effective image alignment, making the process more accessible and efficient. Our code is available at: https://bgu-cs-vil.github.io/SpaceJAM/.
Authors: Gilhan Park, WonJun Moon, SuBeen Lee, Tae-Young Kim, Jae-Pil Heo
Abstract: Class-Incremental Semantic Segmentation(CISS) aims to learn new classes without forgetting the old ones, using only the labels of the new classes. To achieve this, two popular strategies are employed: 1) pseudo-labeling and knowledge distillation to preserve prior knowledge; and 2) background weight transfer, which leverages the broad coverage of background in learning new classes by transferring background weight to the new class classifier. However, the first strategy heavily relies on the old model in detecting old classes while undetected pixels are regarded as the background, thereby leading to the background shift towards the old classes(i.e., misclassification of old class as background). Additionally, in the case of the second approach, initializing the new class classifier with background knowledge triggers a similar background shift issue, but towards the new classes. To address these issues, we propose a background-class separation framework for CISS. To begin with, selective pseudo-labeling and adaptive feature distillation are to distill only trustworthy past knowledge. On the other hand, we encourage the separation between the background and new classes with a novel orthogonal objective along with label-guided output distillation. Our state-of-the-art results validate the effectiveness of these proposed methods.
Authors: Amr Ghoneim, Jiju Poovvancheri, Yasushi Akiyama, Dong Chen
Abstract: Image composition is a complex task which requires a lot of information about the scene for an accurate and realistic composition, such as perspective, lighting, shadows, occlusions, and object interactions. Previous methods have predominantly used 2D information for image composition, neglecting the potentials of 3D spatial information. In this work, we propose DepGAN, a Generative Adversarial Network that utilizes depth maps and alpha channels to rectify inaccurate occlusions and enhance transparency effects in image composition. Central to our network is a novel loss function called Depth Aware Loss which quantifies the pixel wise depth difference to accurately delineate occlusion boundaries while compositing objects at different depth levels. Furthermore, we enhance our network's learning process by utilizing opacity data, enabling it to effectively manage compositions involving transparent and semi-transparent objects. We tested our model against state-of-the-art image composition GANs on benchmark (both real and synthetic) datasets. The results reveal that DepGAN significantly outperforms existing methods in terms of accuracy of object placement semantics, transparency and occlusion handling, both visually and quantitatively. Our code is available at https://amrtsg.github.io/DepGAN/.
Authors: Zehan Wang, Ziang Zhang, Hang Zhang, Luping Liu, Rongjie Huang, Xize Cheng, Hengshuang Zhao, Zhou Zhao
Abstract: Recently, human-computer interaction with various modalities has shown promising applications, like GPT-4o and Gemini. Given the foundational role of multimodal joint representation in understanding and generation pipelines, high-quality omni joint representations would be a step toward co-processing more diverse multimodal information. In this work, we present OmniBind, large-scale multimodal joint representation models ranging in scale from 7 billion to 30 billion parameters, which support 3D, audio, image, and language inputs. Due to the scarcity of data pairs across all modalities, instead of training large models from scratch, we propose remapping and binding the spaces of various pre-trained specialist models together. This approach enables "scaling up" by indirectly increasing the model parameters and the amount of seen data. To effectively integrate various spaces, we dynamically assign weights to different spaces by learning routers with two objectives: cross-modal overall alignment and language representation decoupling. Notably, since binding and routing spaces both only require lightweight networks, OmniBind is extremely training-efficient. Learning the largest 30B model requires merely unpaired unimodal data and approximately 3 days on a single 8-4090 node. Extensive experiments demonstrate the versatility and superiority of OmniBind as an omni representation model, highlighting its great potential for diverse applications, such as any-query and composable multimodal understanding.
Authors: Qi Li, Runpeng Yu, Xinchao Wang
Abstract: This paradigm encapsulates knowledge from various models into a solitary prompt without altering the original models or requiring access to the training data, which enables us to achieve efficient and convenient knowledge transfer in more realistic scenarios. From a practicality standpoint, this paradigm not only for the first time proves the effectiveness of Visual Prompt in data inaccessible contexts, but also solves the problems of low model reusability and high storage resource consumption faced by traditional Data-Free Knowledge Transfer, which means that we can realize the parallel knowledge transfer of multiple models without modifying any source model. Extensive experiments across various datasets and models demonstrate the efficacy of the proposed KiOP knowledge transfer paradigm. Without access to real training data and with rigorous storage capacity constraints, it is also capable of yielding considerable outcomes when dealing with cross-model backbone setups and handling parallel knowledge transfer processing requests with multiple (more than 2) models.
Authors: Hao Ding, Tuxun Lu, Yuqian Zhang, Ruixing Liang, Hongchao Shu, Lalithkumar Seenivasan, Yonghao Long, Qi Dou, Cong Gao, Mathias Unberath
Abstract: Accurate segmentation of tools in robot-assisted surgery is critical for machine perception, as it facilitates numerous downstream tasks including augmented reality feedback. While current feed-forward neural network-based methods exhibit excellent segmentation performance under ideal conditions, these models have proven susceptible to even minor corruptions, significantly impairing the model's performance. This vulnerability is especially problematic in surgical settings where predictions might be used to inform high-stakes decisions. To better understand model behavior under non-adversarial corruptions, prior work has explored introducing artificial corruptions, like Gaussian noise or contrast perturbation to test set images, to assess model robustness. However, these corruptions are either not photo-realistic or model/task agnostic. Thus, these investigations provide limited insights into model deterioration under realistic surgical corruptions. To address this limitation, we introduce the SegSTRONG-C challenge that aims to promote the development of algorithms robust to unforeseen but plausible image corruptions of surgery, like smoke, bleeding, and low brightness. We collect and release corruption-free mock endoscopic video sequences for the challenge participants to train their algorithms and benchmark them on video sequences with photo-realistic non-adversarial corruptions for a binary robot tool segmentation task. This new benchmark will allow us to carefully study neural network robustness to non-adversarial corruptions of surgery, thus constituting an important first step towards more robust models for surgical computer vision. In this paper, we describe the data collection and annotation protocol, baseline evaluations of established segmentation models, and data augmentation-based techniques to enhance model robustness.
Authors: Robin Hesse, Simone Schaub-Meyer, Stefan Roth
Abstract: Attribution maps are one of the most established tools to explain the functioning of computer vision models. They assign importance scores to input features, indicating how relevant each feature is for the prediction of a deep neural network. While much research has gone into proposing new attribution methods, their proper evaluation remains a difficult challenge. In this work, we propose a novel evaluation protocol that overcomes two fundamental limitations of the widely used incremental-deletion protocol, i.e., the out-of-domain issue and lacking inter-model comparisons. This allows us to evaluate 23 attribution methods and how eight different design choices of popular vision models affect their attribution quality. We find that intrinsically explainable models outperform standard models and that raw attribution values exhibit a higher attribution quality than what is known from previous work. Further, we show consistent changes in the attribution quality when varying the network design, indicating that some standard design choices promote attribution quality.
Authors: Tim Elsner, Paula Usinger, Victor Czech, Gregor Kobsik, Yanjiang He, Isaak Lim, Leif Kobbelt
Abstract: In quantised autoencoders, images are usually split into local patches, each encoded by one token. This representation is redundant in the sense that the same number of tokens is spend per region, regardless of the visual information content in that region. Adaptive discretisation schemes like quadtrees are applied to allocate tokens for patches with varying sizes, but this just varies the region of influence for a token which nevertheless remains a local descriptor. Modern architectures add an attention mechanism to the autoencoder which infuses some degree of global information into the local tokens. Despite the global context, tokens are still associated with a local image region. In contrast, our method is inspired by spectral decompositions which transform an input signal into a superposition of global frequencies. Taking the data-driven perspective, we learn custom basis functions corresponding to the codebook entries in our VQ-VAE setup. Furthermore, a decoder combines these basis functions in a non-linear fashion, going beyond the simple linear superposition of spectral decompositions. We can achieve this global description with an efficient transpose operation between features and channels and demonstrate our performance on compression.
Authors: Wenxiang Jiang (Ocean University of China), Hanwei Zhang (Saarland University, Institute of Intelligent Software, Guangzhou), Shuo Zhao (Ocean University of China), Zhongwen Guo (Ocean University of China), Hao Wang (Xidian University, China)
Abstract: Neural Radiance Field (NeRF) represents a significant advancement in computer vision, offering implicit neural network-based scene representation and novel view synthesis capabilities. Its applications span diverse fields including robotics, urban mapping, autonomous navigation, virtual reality/augmented reality, etc., some of which are considered high-risk AI applications. However, despite its widespread adoption, the robustness and security of NeRF remain largely unexplored. In this study, we contribute to this area by introducing the Illusory Poisoning Attack against Neural Radiance Fields (IPA-NeRF). This attack involves embedding a hidden backdoor view into NeRF, allowing it to produce predetermined outputs, i.e. illusory, when presented with the specified backdoor view while maintaining normal performance with standard inputs. Our attack is specifically designed to deceive users or downstream models at a particular position while ensuring that any abnormalities in NeRF remain undetectable from other viewpoints. Experimental results demonstrate the effectiveness of our Illusory Poisoning Attack, successfully presenting the desired illusory on the specified viewpoint without impacting other views. Notably, we achieve this attack by introducing small perturbations solely to the training set. The code can be found at https://github.com/jiang-wenxiang/IPA-NeRF.
Authors: Haoyang He, Jiangning Zhang, Guanzhong Tian, Chengjie Wang, Lei Xie
Abstract: This study explores the recently proposed challenging multi-view Anomaly Detection (AD) task. Single-view tasks would encounter blind spots from other perspectives, resulting in inaccuracies in sample-level prediction. Therefore, we introduce the \textbf{M}ulti-\textbf{V}iew \textbf{A}nomaly \textbf{D}etection (\textbf{MVAD}) framework, which learns and integrates features from multi-views. Specifically, we proposed a \textbf{M}ulti-\textbf{V}iew \textbf{A}daptive \textbf{S}election (\textbf{MVAS}) algorithm for feature learning and fusion across multiple views. The feature maps are divided into neighbourhood attention windows to calculate a semantic correlation matrix between single-view windows and all other views, which is a conducted attention mechanism for each single-view window and the top-K most correlated multi-view windows. Adjusting the window sizes and top-K can minimise the computational complexity to linear. Extensive experiments on the Real-IAD dataset for cross-setting (multi/single-class) validate the effectiveness of our approach, achieving state-of-the-art performance among sample \textbf{4.1\%}$\uparrow$/ image \textbf{5.6\%}$\uparrow$/pixel \textbf{6.7\%}$\uparrow$ levels with a total of ten metrics with only \textbf{18M} parameters and fewer GPU memory and training time.
Authors: Kai Del Regno, Alexander Vilesov, Adnan Armouti, Anirudh Bindiganavale Harish, Selim Emir Can, Ashley Kita, Achuta Kadambi
Abstract: Polysomnography (PSG), the current gold standard method for monitoring and detecting sleep disorders, is cumbersome and costly. At-home testing solutions, known as home sleep apnea testing (HSAT), exist. However, they are contact-based, a feature which limits the ability of some patient populations to tolerate testing and discourages widespread deployment. Previous work on non-contact sleep monitoring for sleep apnea detection either estimates respiratory effort using radar or nasal airflow using a thermal camera, but has not compared the two or used them together. We conducted a study on 10 participants, ages 34 - 78, with suspected sleep disorders using a hardware setup with a synchronized radar and thermal camera. We show the first comparison of radar and thermal imaging for sleep monitoring, and find that our thermal imaging method outperforms radar significantly. Our thermal imaging method detects apneas with an accuracy of 0.99, a precision of 0.68, a recall of 0.74, an F1 score of 0.71, and an intra-class correlation of 0.70; our radar method detects apneas with an accuracy of 0.83, a precision of 0.13, a recall of 0.86, an F1 score of 0.22, and an intra-class correlation of 0.13. We also present a novel proposal for classifying obstructive and central sleep apnea by leveraging a multimodal setup. This method could be used accurately detect and classify apneas during sleep with non-contact sensors, thereby improving diagnostic capacities in patient populations unable to tolerate current technology.
Authors: Marco Huber, Naser Damer
Abstract: The need for more transparent face recognition (FR), along with other visual-based decision-making systems has recently attracted more attention in research, society, and industry. The reasons why two face images are matched or not matched by a deep learning-based face recognition system are not obvious due to the high number of parameters and the complexity of the models. However, it is important for users, operators, and developers to ensure trust and accountability of the system and to analyze drawbacks such as biased behavior. While many previous works use spatial semantic maps to highlight the regions that have a significant influence on the decision of the face recognition system, frequency components which are also considered by CNNs, are neglected. In this work, we take a step forward and investigate explainable face recognition in the unexplored frequency domain. This makes this work the first to propose explainability of verification-based decisions in the frequency domain, thus explaining the relative influence of the frequency components of each input toward the obtained outcome. To achieve this, we manipulate face images in the spatial frequency domain and investigate the impact on verification outcomes. In extensive quantitative experiments, along with investigating two special scenarios cases, cross-resolution FR and morphing attacks (the latter in supplementary material), we observe the applicability of our proposed frequency-based explanations.
Authors: Ping Wang, Yulun Zhang, Lishun Wang, Xin Yuan
Abstract: Transformers have achieved the state-of-the-art performance on solving the inverse problem of Snapshot Compressive Imaging (SCI) for video, whose ill-posedness is rooted in the mixed degradation of spatial masking and temporal aliasing. However, previous Transformers lack an insight into the degradation and thus have limited performance and efficiency. In this work, we tailor an efficient reconstruction architecture without temporal aggregation in early layers and Hierarchical Separable Video Transformer (HiSViT) as building block. HiSViT is built by multiple groups of Cross-Scale Separable Multi-head Self-Attention (CSS-MSA) and Gated Self-Modulated Feed-Forward Network (GSM-FFN) with dense connections, each of which is conducted within a separate channel portions at a different scale, for multi-scale interactions and long-range modeling. By separating spatial operations from temporal ones, CSS-MSA introduces an inductive bias of paying more attention within frames instead of between frames while saving computational overheads. GSM-FFN is design to enhance the locality via gated mechanism and factorized spatial-temporal convolutions. Extensive experiments demonstrate that our method outperforms previous methods by $>\!0.5$ dB with comparable or fewer complexity and parameters. The source codes and pretrained models are released at https://github.com/pwangcs/HiSViT.
Authors: Jiaxi Zeng, Chengtang Yao, Yuwei Wu, Yunde Jia
Abstract: Stereo matching provides depth estimation from binocular images for downstream applications. These applications mostly take video streams as input and require temporally consistent depth maps. However, existing methods mainly focus on the estimation at the single-frame level. This commonly leads to temporally inconsistent results, especially in ill-posed regions. In this paper, we aim to leverage temporal information to improve the temporal consistency, accuracy, and efficiency of stereo matching. To achieve this, we formulate video stereo matching as a process of temporal disparity completion followed by continuous iterative refinements. Specifically, we first project the disparity of the previous timestamp to the current viewpoint, obtaining a semi-dense disparity map. Then, we complete this map through a disparity completion module to obtain a well-initialized disparity map. The state features from the current completion module and from the past refinement are fused together, providing a temporally coherent state for subsequent refinement. Based on this coherent state, we introduce a dual-space refinement module to iteratively refine the initialized result in both disparity and disparity gradient spaces, improving estimations in ill-posed regions. Extensive experiments demonstrate that our method effectively alleviates temporal inconsistency while enhancing both accuracy and efficiency.
Authors: Olga Zatsarynna, Emad Bahrami, Yazan Abu Farha, Gianpiero Francesca, Juergen Gall
Abstract: Long-term action anticipation has become an important task for many applications such as autonomous driving and human-robot interaction. Unlike short-term anticipation, predicting more actions into the future imposes a real challenge with the increasing uncertainty in longer horizons. While there has been a significant progress in predicting more actions into the future, most of the proposed methods address the task in a deterministic setup and ignore the underlying uncertainty. In this paper, we propose a novel Gated Temporal Diffusion (GTD) network that models the uncertainty of both the observation and the future predictions. As generator, we introduce a Gated Anticipation Network (GTAN) to model both observed and unobserved frames of a video in a mutual representation. On the one hand, using a mutual representation for past and future allows us to jointly model ambiguities in the observation and future, while on the other hand GTAN can by design treat the observed and unobserved parts differently and steer the information flow between them. Our model achieves state-of-the-art results on the Breakfast, Assembly101 and 50Salads datasets in both stochastic and deterministic settings. Code: https://github.com/olga-zats/GTDA .
Authors: Jaehyeok Kim, Dongyoon Wee, Dan Xu
Abstract: This paper introduces Motion-oriented Compositional Neural Radiance Fields (MoCo-NeRF), a framework designed to perform free-viewpoint rendering of monocular human videos via novel non-rigid motion modeling approach. In the context of dynamic clothed humans, complex cloth dynamics generate non-rigid motions that are intrinsically distinct from skeletal articulations and critically important for the rendering quality. The conventional approach models non-rigid motions as spatial (3D) deviations in addition to skeletal transformations. However, it is either time-consuming or challenging to achieve optimal quality due to its high learning complexity without a direct supervision. To target this problem, we propose a novel approach of modeling non-rigid motions as radiance residual fields to benefit from more direct color supervision in the rendering and utilize the rigid radiance fields as a prior to reduce the complexity of the learning process. Our approach utilizes a single multiresolution hash encoding (MHE) to concurrently learn the canonical T-pose representation from rigid skeletal motions and the radiance residual field for non-rigid motions. Additionally, to further improve both training efficiency and usability, we extend MoCo-NeRF to support simultaneous training of multiple subjects within a single framework, thanks to our effective design for modeling non-rigid motions. This scalability is achieved through the integration of a global MHE and learnable identity codes in addition to multiple local MHEs. We present extensive results on ZJU-MoCap and MonoCap, clearly demonstrating state-of-the-art performance in both single- and multi-subject settings. The code and model will be made publicly available at the project page: https://stevejaehyeok.github.io/publications/moco-nerf.
URLs: https://stevejaehyeok.github.io/publications/moco-nerf.
Authors: Yu Shang, Jiansheng Chen, Hangyu Fan, Jingtao Ding, Jie Feng, Yong Li
Abstract: Cities, as the most fundamental environment of human life, encompass diverse physical elements such as buildings, roads and vegetation with complex interconnection. Crafting realistic, interactive 3D urban environments plays a crucial role in constructing AI agents capable of perceiving, decision-making, and acting like humans in real-world environments. However, creating high-fidelity 3D urban environments usually entails extensive manual labor from designers, involving intricate detailing and accurate representation of complex urban features. Therefore, how to accomplish this in an automatical way remains a longstanding challenge. Toward this problem, we propose UrbanWorld, the first generative urban world model that can automatically create a customized, realistic and interactive 3D urban world with flexible control conditions. UrbanWorld incorporates four key stages in the automatical crafting pipeline: 3D layout generation from openly accessible OSM data, urban scene planning and designing with a powerful urban multimodal large language model (Urban MLLM), controllable urban asset rendering with advanced 3D diffusion techniques, and finally the MLLM-assisted scene refinement. The crafted high-fidelity 3D urban environments enable realistic feedback and interactions for general AI and machine perceptual systems in simulations. We are working on contributing UrbanWorld as an open-source and versatile platform for evaluating and improving AI abilities in perception, decision-making, and interaction in realistic urban environments.
Authors: Yifan Gong, Zheng Zhan, Yanyu Li, Yerlan Idelbayev, Andrey Zharkov, Kfir Aberman, Sergey Tulyakov, Yanzhi Wang, Jian Ren
Abstract: Good weight initialization serves as an effective measure to reduce the training cost of a deep neural network (DNN) model. The choice of how to initialize parameters is challenging and may require manual tuning, which can be time-consuming and prone to human error. To overcome such limitations, this work takes a novel step towards building a weight generator to synthesize the neural weights for initialization. We use the image-to-image translation task with generative adversarial networks (GANs) as an example due to the ease of collecting model weights spanning a wide range. Specifically, we first collect a dataset with various image editing concepts and their corresponding trained weights, which are later used for the training of the weight generator. To address the different characteristics among layers and the substantial number of weights to be predicted, we divide the weights into equal-sized blocks and assign each block an index. Subsequently, a diffusion model is trained with such a dataset using both text conditions of the concept and the block indexes. By initializing the image translation model with the denoised weights predicted by our diffusion model, the training requires only 43.3 seconds. Compared to training from scratch (i.e., Pix2pix), we achieve a 15x training time acceleration for a new concept while obtaining even better image generation quality.
Authors: Stanley Cao, Kevin Liu
Abstract: Figures are essential channels for densely communicating complex ideas in scientific papers. Previous work in automatically generating figure captions has been largely unsuccessful and has defaulted to using single-layer LSTMs, which no longer achieve state-of-the-art performance. In our work, we use the SciCap datasets curated by Hsu et al. and use a variant of a CLIP+GPT-2 encoder-decoder model with cross-attention to generate captions conditioned on the image. Furthermore, we augment our training pipeline by creating a new dataset MetaSciCap that incorporates textual metadata from the original paper relevant to the figure, such as the title, abstract, and in-text references. We use SciBERT to encode the textual metadata and use this encoding alongside the figure embedding. In our experimentation with different models, we found that the CLIP+GPT-2 model performs better when it receives all textual metadata from the SciBERT encoder in addition to the figure, but employing a SciBERT+GPT2 model that uses only the textual metadata achieved optimal performance.
Authors: Yifan Zhu, Yibo Miao, Yinpeng Dong, Xiao-Shan Gao
Abstract: Despite the great progress of 3D vision, data privacy and security issues in 3D deep learning are not explored systematically. In the domain of 2D images, many availability attacks have been proposed to prevent data from being illicitly learned by unauthorized deep models. However, unlike images represented on a fixed dimensional grid, point clouds are characterized as unordered and unstructured sets, posing a significant challenge in designing an effective availability attack for 3D deep learning. In this paper, we theoretically show that extending 2D availability attacks directly to 3D point clouds under distance regularization is susceptible to the degeneracy, rendering the generated poisons weaker or even ineffective. This is because in bi-level optimization, introducing regularization term can result in update directions out of control. To address this issue, we propose a novel Feature Collision Error-Minimization (FC-EM) method, which creates additional shortcuts in the feature space, inducing different update directions to prevent the degeneracy of bi-level optimization. Moreover, we provide a theoretical analysis that demonstrates the effectiveness of the FC-EM attack. Extensive experiments on typical point cloud datasets, 3D intracranial aneurysm medical dataset, and 3D face dataset verify the superiority and practicality of our approach. Code is available at https://github.com/hala64/fc-em.
Authors: Ivan S. Maksymov
Abstract: The discovery of the quantum tunnelling effect -- the transmission of particles through a high potential barrier -- was one of the most impressive achievements of quantum mechanics made in the 1920s. Responding to the contemporary challenges, I introduce a novel deep neural network (DNN) architecture that processes information using the effect of quantum tunnelling. I demonstrate the ability of the quantum tunnelling DNN (QT-DNN) to recognise optical illusions like a human. Hardware implementation of QT-DNN is expected to result in an inexpensive and energy-efficient neuromorphic chip suitable for applications in autonomous vehicles. The optical illusions recognition tests developed in this paper should lay foundations for cognitive benchmarking tasks for AI systems of the future, benefiting the fields of sociophysics and behavioural science.
Authors: Dezan Zhao
Abstract: The model quantization technique of deep neural networks has garnered significant attention and has proven to be highly useful in compressing model size, reducing computation costs, and accelerating inference. Many researchers employ fake quantization for analyzing or training the quantization process. However, fake quantization is not the final form for deployment, and there exists a gap between the academic setting and real-world deployment. Additionally, the inclusion of additional computation with scale and zero-point makes deployment a challenging task. In this study, we first analyze why the combination of quantization and dequantization is used to train the model and draw the conclusion that fake quantization research is reasonable due to the disappearance of weight gradients and the ability to approximate between fake and real quantization. Secondly, we propose ISQuant as a solution for deploying 8-bit models. ISQuant is fast and easy to use for most 8-bit models, requiring fewer parameters and less computation. ISQuant also inherits the advantages of SQuant, such as not requiring training data and being very fast at the first level of quantization. Finally We conduct some experiments and found the results is acceptable.our code is available at https://github.com/
URLs: https://github.com/
Authors: Mingyuan Fan, Yang Liu, Cen Chen, Ximeng Liu
Abstract: Adversarial attack has garnered considerable attention due to its profound implications for the secure deployment of robots in sensitive security scenarios. To potentially push for advances in the field, this paper studies the adversarial attack in the black-box setting and proposes an unlabeled data-driven adversarial attack method, called SemiAdv. Specifically, SemiAdv achieves the following breakthroughs compared with previous works. First, by introducing the semi-supervised learning technique into the adversarial attack, SemiAdv substantially decreases the number of queries required for generating adversarial samples. On average, SemiAdv only needs to query a few hundred times to launch an effective attack with more than 90% success rate. Second, many existing black-box adversarial attacks require massive labeled data to mitigate the difference between the local substitute model and the remote target model for a good attack performance. While SemiAdv relaxes this limitation and is capable of utilizing unlabeled raw data to launch an effective attack. Finally, our experiments show that SemiAdv saves up to 12x query accesses for generating adversarial samples while maintaining a competitive attack success rate compared with state-of-the-art attacks.
Authors: Thinh Nguyen, Khoa D Doan, Binh T. Nguyen, Danh Le-Phuoc, Kok-Seng Wong
Abstract: Federated Class-Incremental Learning (FCIL) increasingly becomes important in the decentralized setting, where it enables multiple participants to collaboratively train a global model to perform well on a sequence of tasks without sharing their private data. In FCIL, conventional Federated Learning algorithms such as FedAVG often suffer from catastrophic forgetting, resulting in significant performance declines on earlier tasks. Recent works, based on generative models, produce synthetic images to help mitigate this issue across all classes, but these approaches' testing accuracy on previous classes is still much lower than recent classes, i.e., having better plasticity than stability. To overcome these issues, this paper presents Federated Global Twin Generator (FedGTG), an FCIL framework that exploits privacy-preserving generative-model training on the global side without accessing client data. Specifically, the server trains a data generator and a feature generator to create two types of information from all seen classes, and then it sends the synthetic data to the client side. The clients then use feature-direction-controlling losses to make the local models retain knowledge and learn new tasks well. We extensively analyze the robustness of FedGTG on natural images, as well as its ability to converge to flat local minima and achieve better-predicting confidence (calibration). Experimental results on CIFAR-10, CIFAR-100, and tiny-ImageNet demonstrate the improvements in accuracy and forgetting measures of FedGTG compared to previous frameworks.
Authors: Maohan Liang, Ryan Wen Liu, Ruobin Gao, Zhe Xiao, Xiaocai Zhang, Hua Wang
Abstract: Vessel trajectory clustering, a crucial component of the maritime intelligent transportation systems, provides valuable insights for applications such as anomaly detection and trajectory prediction. This paper presents a comprehensive survey of the most prevalent distance-based vessel trajectory clustering methods, which encompass two main steps: trajectory similarity measurement and clustering. Initially, we conducted a thorough literature review using relevant keywords to gather and summarize pertinent research papers and datasets. Then, this paper discussed the principal methods of data pre-processing that prepare data for further analysis. The survey progresses to detail the leading algorithms for measuring vessel trajectory similarity and the main clustering techniques used in the field today. Furthermore, the various applications of trajectory clustering within the maritime context are explored. Finally, the paper evaluates the effectiveness of different algorithm combinations and pre-processing methods through experimental analysis, focusing on their impact on the performance of distance-based trajectory clustering algorithms. The experimental results demonstrate the effectiveness of various trajectory clustering algorithms and notably highlight the significant improvements that trajectory compression techniques contribute to the efficiency and accuracy of trajectory clustering. This comprehensive approach ensures a deep understanding of current capabilities and future directions in vessel trajectory clustering.
Authors: Zhiwen Yang, Hui Zhang, Dan Zhao, Bingzheng Wei, Yan Xu
Abstract: Transformers have revolutionized medical image restoration, but the quadratic complexity still poses limitations for their application to high-resolution medical images. The recent advent of RWKV in the NLP field has attracted much attention as it can process long sequences efficiently. To leverage its advanced design, we propose Restore-RWKV, the first RWKV-based model for medical image restoration. Since the original RWKV model is designed for 1D sequences, we make two necessary modifications for modeling spatial relations in 2D images. First, we present a recurrent WKV (Re-WKV) attention mechanism that captures global dependencies with linear computational complexity. Re-WKV incorporates bidirectional attention as basic for a global receptive field and recurrent attention to effectively model 2D dependencies from various scan directions. Second, we develop an omnidirectional token shift (Omni-Shift) layer that enhances local dependencies by shifting tokens from all directions and across a wide context range. These adaptations make the proposed Restore-RWKV an efficient and effective model for medical image restoration. Extensive experiments demonstrate that Restore-RWKV achieves superior performance across various medical image restoration tasks, including MRI image super-resolution, CT image denoising, PET image synthesis, and all-in-one medical image restoration. Code is available at: \href{https://github.com/Yaziwel/Restore-RWKV.git}{https://github.com/Yaziwel/Restore-RWKV}.
URLs: https://github.com/Yaziwel/Restore-RWKV.git, https://github.com/Yaziwel/Restore-RWKV
Authors: Danish Nazir, Timo Bartels, Jan Piewek, Thorsten Bagdonat, Tim Fingscheidt
Abstract: Distributed computing in the context of deep neural networks (DNNs) implies the execution of one part of the network on edge devices and the other part typically on a large-scale cloud platform. Conventional methods propose to employ a serial concatenation of a learned image and source encoder, the latter projecting the image encoder output (bottleneck features) into a quantized representation for bitrate-efficient transmission. In the cloud, a respective source decoder reprojects the quantized representation to the original feature representation, serving as an input for the downstream task decoder performing, e.g., semantic segmentation. In this work, we propose joint source and task decoding, as it allows for a smaller network size in the cloud. This further enables the scalability of such services in large numbers without requiring extensive computational load on the cloud per channel. We demonstrate the effectiveness of our method by achieving a distributed semantic segmentation SOTA over a wide range of bitrates on the mean intersection over union metric, while using only $9.8 \%$ ... $11.59 \%$ of cloud DNN parameters used in the previous SOTA on the COCO and Cityscapes datasets.
Authors: Srija Mukhopadhyay, Adnan Qidwai, Aparna Garimella, Pritika Ramu, Vivek Gupta, Dan Roth
Abstract: Chart question answering (CQA) is a crucial area of Visual Language Understanding. However, the robustness and consistency of current Visual Language Models (VLMs) in this field remain under-explored. This paper evaluates state-of-the-art VLMs on comprehensive datasets, developed specifically for this study, encompassing diverse question categories and chart formats. We investigate two key aspects: 1) the models' ability to handle varying levels of chart and question complexity, and 2) their robustness across different visual representations of the same underlying data. Our analysis reveals significant performance variations based on question and chart types, highlighting both strengths and weaknesses of current models. Additionally, we identify areas for improvement and propose future research directions to build more robust and reliable CQA systems. This study sheds light on the limitations of current models and paves the way for future advancements in the field.
Authors: Ya\c{s}ar Utku Al\c{c}alar, Mehmet Ak\c{c}akaya
Abstract: Diffusion models have emerged as powerful generative techniques for solving inverse problems. Despite their success in a variety of inverse problems in imaging, these models require many steps to converge, leading to slow inference time. Recently, there has been a trend in diffusion models for employing sophisticated noise schedules that involve more frequent iterations of timesteps at lower noise levels, thereby improving image generation and convergence speed. However, application of these ideas for solving inverse problems with diffusion models remain challenging, as these noise schedules do not perform well when using empirical tuning for the forward model log-likelihood term weights. To tackle these challenges, we propose zero-shot approximate posterior sampling (ZAPS) that leverages connections to zero-shot physics-driven deep learning. ZAPS fixes the number of sampling steps, and uses zero-shot training with a physics-guided loss function to learn log-likelihood weights at each irregular timestep. We apply ZAPS to the recently proposed diffusion posterior sampling method as baseline, though ZAPS can also be used with other posterior sampling diffusion models. We further approximate the Hessian of the logarithm of the prior using a diagonalization approach with learnable diagonal entries for computational efficiency. These parameters are optimized over a fixed number of epochs with a given computational budget. Our results for various noisy inverse problems, including Gaussian and motion deblurring, inpainting, and super-resolution show that ZAPS reduces inference time, provides robustness to irregular noise schedules and improves reconstruction quality. Code is available at https://github.com/ualcalar17/ZAPS
Authors: Jiantao Feng, Xinde Li, HyunCheol Park, Juan Liu, Zhentong Zhang
Abstract: Simultaneous Localization and Mapping (SLAM) technology enables the construction of environmental maps and localization, serving as a key technique for indoor autonomous navigation of mobile robots. Traditional SLAM methods typically require exhaustive traversal of all rooms during indoor navigation to obtain a complete map, resulting in lengthy path planning times and prolonged time to reach target points. Moreover, cumulative errors during motion lead to inaccurate robot localization, impacting navigation efficiency.This paper proposes an improved SLAM method, FR-SLAM, based on floor plan registration, utilizing a morphology-based floor plan registration algorithm to align and transform original floor plans. This approach facilitates the rapid acquisition of comprehensive motion maps and efficient path planning, enabling swift navigation to target positions within a shorter timeframe. To enhance registration and robot motion localization accuracy, a real-time update strategy is employed, comparing the current position's building structure with the map and dynamically updating floor plan registration results for precise localization. Comparative tests conducted on real and simulated datasets demonstrate that, compared to other benchmark algorithms, this method achieves higher floor plan registration accuracy and shorter time consumption to reach target positions.
Authors: Arianna Bunnell, Kailee Hung, John A. Shepherd, Peter Sadowski
Abstract: Development of artificial intelligence (AI) for medical imaging demands curation and cleaning of large-scale clinical datasets comprising hundreds of thousands of images. Some modalities, such as mammography, contain highly standardized imaging. In contrast, breast ultrasound imaging (BUS) can contain many irregularities not indicated by scan metadata, such as enhanced scan modes, sonographer annotations, or additional views. We present an open-source software solution for automatically processing clinical BUS datasets. The algorithm performs BUS scan filtering, cleaning, and knowledge extraction from sonographer annotations. Its modular design enables users to adapt it to new settings. Experiments on an internal testing dataset of 430 clinical BUS images achieve >95% sensitivity and >98% specificity in detecting every type of text annotation, >98% sensitivity and specificity in detecting scans with blood flow highlighting, alternative scan modes, or invalid scans. A case study on a completely external, public dataset of BUS scans found that BUSClean identified text annotations and scans with blood flow highlighting with 88.6% and 90.9% sensitivity and 98.3% and 99.9% specificity, respectively. Adaptation of the lesion caliper detection method to account for a type of caliper specific to the case study demonstrates intended use of BUSClean in new data distributions and improved performance in lesion caliper detection from 43.3% and 93.3% out-of-the-box to 92.1% and 92.3% sensitivity and specificity, respectively. Source code, example notebooks, and sample data are available at https://github.com/hawaii-ai/bus-cleaning.
Authors: Siyuan Cheng, Guangyu Shen, Kaiyuan Zhang, Guanhong Tao, Shengwei An, Hanxi Guo, Shiqing Ma, Xiangyu Zhang
Abstract: Deep neural networks (DNNs) have demonstrated effectiveness in various fields. However, DNNs are vulnerable to backdoor attacks, which inject a unique pattern, called trigger, into the input to cause misclassification to an attack-chosen target label. While existing works have proposed various methods to mitigate backdoor effects in poisoned models, they tend to be less effective against recent advanced attacks. In this paper, we introduce a novel post-training defense technique UNIT that can effectively eliminate backdoor effects for a variety of attacks. In specific, UNIT approximates a unique and tight activation distribution for each neuron in the model. It then proactively dispels substantially large activation values that exceed the approximated boundaries. Our experimental results demonstrate that UNIT outperforms 7 popular defense methods against 14 existing backdoor attacks, including 2 advanced attacks, using only 5\% of clean training data. UNIT is also cost efficient. The code is accessible at https://github.com/Megum1/UNIT.
Authors: Guobiao Li, Sheng Li, Zhenxing Qian, Xinpeng Zhang
Abstract: Image steganography is the process of hiding secret data in a cover image by subtle perturbation. Recent studies show that it is feasible to use a fixed neural network for data embedding and extraction. Such Fixed Neural Network Steganography (FNNS) demonstrates favorable performance without the need for training networks, making it more practical for real-world applications. However, the stego-images generated by the existing FNNS methods exhibit high distortion, which is prone to be detected by steganalysis tools. To deal with this issue, we propose a Cover-separable Fixed Neural Network Steganography, namely Cs-FNNS. In Cs-FNNS, we propose a Steganographic Perturbation Search (SPS) algorithm to directly encode the secret data into an imperceptible perturbation, which is combined with an AI-generated cover image for transmission. Through accessing the same deep generative models, the receiver could reproduce the cover image using a pre-agreed key, to separate the perturbation in the stego-image for data decoding. such an encoding/decoding strategy focuses on the secret data and eliminates the disturbance of the cover images, hence achieving a better performance. We apply our Cs-FNNS to the steganographic field that hiding secret images within cover images. Through comprehensive experiments, we demonstrate the superior performance of the proposed method in terms of visual quality and undetectability. Moreover, we show the flexibility of our Cs-FNNS in terms of hiding multiple secret images for different receivers.
Authors: Jaehoon Hahm, Junho Lee, Sunghyun Kim, Joonseok Lee
Abstract: The latent space of diffusion model mostly still remains unexplored, despite its great success and potential in the field of generative modeling. In fact, the latent space of existing diffusion models are entangled, with a distorted mapping from its latent space to image space. To tackle this problem, we present Isometric Diffusion, equipping a diffusion model with a geometric regularizer to guide the model to learn a geometrically sound latent space of the training data manifold. This approach allows diffusion models to learn a more disentangled latent space, which enables smoother interpolation, more accurate inversion, and more precise control over attributes directly in the latent space. Our extensive experiments consisting of image interpolations, image inversions, and linear editing show the effectiveness of our method.
Authors: Xinyi Wang, Angeliki Katsenou, David Bull
Abstract: With the rapid growth of User-Generated Content (UGC) exchanged between users and sharing platforms, the need for video quality assessment in the wild has emerged. UGC is mostly acquired using consumer devices and undergoes multiple rounds of compression or transcoding before reaching the end user. Therefore, traditional quality metrics that require the original content as a reference cannot be used. In this paper, we propose ReLaX-VQA, a novel No-Reference Video Quality Assessment (NR-VQA) model that aims to address the challenges of evaluating the diversity of video content and the assessment of its quality without reference videos. ReLaX-VQA uses fragments of residual frames and optical flow, along with different expressions of spatial features of the sampled frames, to enhance motion and spatial perception. Furthermore, the model enhances abstraction by employing layer-stacking techniques in deep neural network features (from Residual Networks and Vision Transformers). Extensive testing on four UGC datasets confirms that ReLaX-VQA outperforms existing NR-VQA methods with an average SRCC value of 0.8658 and PLCC value of 0.8872. We will open source the code and trained models to facilitate further research and applications of NR-VQA: https://github.com/xinyiW915/ReLaX-VQA.
Authors: Niamh Belton, Aonghus Lawlor, Kathleen M. Curran
Abstract: The diagnostic accuracy and subjectivity of existing Knee Osteoarthritis (OA) ordinal grading systems has been a subject of on-going debate and concern. Existing automated solutions are trained to emulate these imperfect systems, whilst also being reliant on large annotated databases for fully-supervised training. This work proposes a three stage approach for automated continuous grading of knee OA that is built upon the principles of Anomaly Detection (AD); learning a robust representation of healthy knee X-rays and grading disease severity based on its distance to the centre of normality. In the first stage, SS-FewSOME is proposed, a self-supervised AD technique that learns the 'normal' representation, requiring only examples of healthy subjects and <3% of the labels that existing methods require. In the second stage, this model is used to pseudo label a subset of unlabelled data as 'normal' or 'anomalous', followed by denoising of pseudo labels with CLIP. The final stage involves retraining on labelled and pseudo labelled data using the proposed Dual Centre Representation Learning (DCRL) which learns the centres of two representation spaces; normal and anomalous. Disease severity is then graded based on the distance to the learned centres. The proposed methodology outperforms existing techniques by margins of up to 24% in terms of OA detection and the disease severity scores correlate with the Kellgren-Lawrence grading system at the same level as human expert performance. Code available at https://github.com/niamhbelton/SS-FewSOME_Disease_Severity_Knee_Osteoarthritis.
URLs: https://github.com/niamhbelton/SS-FewSOME_Disease_Severity_Knee_Osteoarthritis.
Authors: Bizhe Bai, Yan-Jie Zhou, Yujian Hu, Tony C. W. Mok, Yilang Xiang, Le Lu, Hongkun Zhang, Minfeng Xu
Abstract: Pulmonary embolism (PE) is a life-threatening condition where rapid and accurate diagnosis is imperative yet difficult due to predominantly atypical symptomatology. Computed tomography pulmonary angiography (CTPA) is acknowledged as the gold standard imaging tool in clinics, yet it can be contraindicated for emergency department (ED) patients and represents an onerous procedure, thus necessitating PE identification through non-contrast CT (NCT) scans. In this work, we explore the feasibility of applying a deep-learning approach to NCT scans for PE identification. We propose a novel Cross-Phase Mutual learNing framework (CPMN) that fosters knowledge transfer from CTPA to NCT, while concurrently conducting embolism segmentation and abnormality classification in a multi-task manner. The proposed CPMN leverages the Inter-Feature Alignment (IFA) strategy that enhances spatial contiguity and mutual learning between the dual-pathway network, while the Intra-Feature Discrepancy (IFD) strategy can facilitate precise segmentation of PE against complex backgrounds for single-pathway networks. For a comprehensive assessment of the proposed approach, a large-scale dual-phase dataset containing 334 PE patients and 1,105 normal subjects has been established. Experimental results demonstrate that CPMN achieves the leading identification performance, which is 95.4\% and 99.6\% in patient-level sensitivity and specificity on NCT scans, indicating the potential of our approach as an economical, accessible, and precise tool for PE identification in clinical practice.
Authors: Zhuoyuan Li, Yao Li, Chuanbo Tang, Li Li, Dong Liu, Feng Wu
Abstract: Inter prediction is a key technology to reduce the temporal redundancy in video coding. In natural videos, there are usually multiple moving objects with variable velocity, resulting in complex motion fields that are difficult to represent compactly. In Versatile Video Coding (VVC), existing inter prediction methods usually assume uniform speed motion between consecutive frames and use the linear models for motion estimation (ME) and motion compensation (MC), which may not well handle the complex motion fields in the real world. To address these issues, we introduce a uniformly accelerated motion model (UAMM) to exploit motion-related elements (velocity, acceleration) of moving objects between the video frames, and further combine them to assist the inter prediction methods to handle the variable motion in the temporal domain. Specifically, first, the theory of UAMM is mentioned. Second, based on that, we propose the UAMM-based parameter derivation and extrapolation schemes in the coding process. Third, we integrate the UAMM into existing inter prediction modes (Merge, MMVD, CIIP) to achieve higher prediction accuracy. The proposed method is implemented into the VVC reference software, VTM version 12.0. Experimental results show that the proposed method achieves up to 0.38% and on average 0.13% BD-rate reduction compared to the VTM anchor, under the Low-delay P configuration, with a slight increase of time complexity on the encoding/decoding side.
Authors: Jixiang Luo
Abstract: Since LIC has made rapid progress recently compared to traditional methods, this paper attempts to discuss the question about 'Where is the boundary of Learned Image Compression(LIC)?' with regard to subjective matrics. Thus this paper splits the above problem into two sub-problems:1)Where is the boundary of rate-distortion performance of PSNR? 2)How to further improve the compression gain and achieve the boundary? Therefore this paper analyzes the effectiveness of scaling parameters for encoder, decoder and context model, which are the three components of LIC. Then we conclude that scaling for LIC is to scale for context model and decoder within LIC. Extensive experiments demonstrate that overfitting can actually serve as an effective context. By optimizing the context, this paper further improves PSNR and achieves state-of-the-art performance, showing a performance gain of 14.39% with BD-RATE over VVC.
Authors: Piotr Mirowski, David Warde-Farley, Mihaela Rosca, Matthew Koichi Grimes, Yana Hasson, Hyunjik Kim, M\'elanie Rey, Simon Osindero, Suman Ravuri, Shakir Mohamed
Abstract: Atmospheric states derived from reanalysis comprise a substantial portion of weather and climate simulation outputs. Many stakeholders -- such as researchers, policy makers, and insurers -- use this data to better understand the earth system and guide policy decisions. Atmospheric states have also received increased interest as machine learning approaches to weather prediction have shown promising results. A key issue for all audiences is that dense time series of these high-dimensional states comprise an enormous amount of data, precluding all but the most well resourced groups from accessing and using historical data and future projections. To address this problem, we propose a method for compressing atmospheric states using methods from the neural network literature, adapting spherical data to processing by conventional neural architectures through the use of the area-preserving HEALPix projection. We investigate two model classes for building neural compressors: the hyperprior model from the neural image compression literature and recent vector-quantised models. We show that both families of models satisfy the desiderata of small average error, a small number of high-error reconstructed pixels, faithful reproduction of extreme events such as hurricanes and heatwaves, preservation of the spectral power distribution across spatial scales. We demonstrate compression ratios in excess of 1000x, with compression and decompression at a rate of approximately one second per global atmospheric state.
Authors: Pedro Felzenszwalb
Abstract: Deconvolution with a box (square wave) is a key operation for super-resolution with pixel-shift cameras. In general convolution with a box is not invertible. However, we can obtain perfect reconstructions of sparse signals using convex optimization. We give a direct proof that improves on the reconstruction bound that follows from previous results. We also show our bound is tight and matches an information theoretic limit.
Authors: Alberto Pirillo, Luca Colombo, Manuel Roveri
Abstract: Quantization has become increasingly pivotal in addressing the steadily increasing computational and memory requirements of Deep Neural Networks (DNNs). By reducing the number of bits used to represent weights and activations (typically from 32-bit floating-point to 16-bit or 8-bit integers), quantization reduces the memory footprint, energy consumption, and execution time of DNN models. However, traditional quantization methods typically focus on the inference of DNNs, while the training process still relies on floating-point operations. To date, only one work in the literature has addressed integer-only training for Multi-Layer Perceptron (MLP) architectures. This work introduces NITRO-D, a new framework for training arbitrarily deep integer-only Convolutional Neural Networks (CNNs) that operate entirely< in the integer-only domain for both training and inference. NITRO-D is the first framework in the literature enabling the training of integer-only CNNs without the need to introduce a quantization scheme. Specifically, NITRO-D introduces a novel architecture integrating multiple integer local-loss blocks, which include the proposed NITRO Scaling Layer and the NITRO-ReLU activation function. Additionally, it introduces a novel integer-only learning algorithm derived from Local Error Signals (LES), utilizing IntegerSGD, an optimizer specifically designed to operate in an integer-only context. NITRO-D is implemented in an open-source Python library. Extensive experimental evaluations demonstrate its effectiveness across several state-of-the-art image recognition datasets. Results show significant performance improvements from 2.47% to 5.96% for integer-only MLP architectures over the state-of-the-art solution, and the capability of training integer-only CNN architectures with minimal accuracy degradation from -0.15% to -4.22% compared to floating-point LES.
Authors: Erik Wallin, Lennart Svensson, Fredrik Kahl, Lars Hammarstrand
Abstract: In open-set semi-supervised learning (OSSL), we consider unlabeled datasets that may contain unknown classes. Existing OSSL methods often use the softmax confidence for classifying data as in-distribution (ID) or out-of-distribution (OOD). Additionally, many works for OSSL rely on ad-hoc thresholds for ID/OOD classification, without considering the statistics of the problem. We propose a new score for ID/OOD classification based on angles in feature space between data and an ID subspace. Moreover, we propose an approach to estimate the conditional distributions of scores given ID or OOD data, enabling probabilistic predictions of data being ID or OOD. These components are put together in a framework for OSSL, termed \emph{ProSub}, that is experimentally shown to reach SOTA performance on several benchmark problems. Our code is available at https://github.com/walline/prosub.
Authors: Jingwen Yu, Hanjing Ye, Jianhao Jiao, Ping Tan, Hong Zhang
Abstract: Visual loop closure detection is an important module in visual simultaneous localization and mapping (SLAM), which associates current camera observation with previously visited places. Loop closures correct drifts in trajectory estimation to build a globally consistent map. However, a false loop closure can be fatal, so verification is required as an additional step to ensure robustness by rejecting the false positive loops. Geometric verification has been a well-acknowledged solution that leverages spatial clues provided by local feature matching to find true positives. Existing feature matching methods focus on homography and pose estimation in long-term visual localization, lacking references for geometric verification. To fill the gap, this paper proposes a unified benchmark targeting geometric verification of loop closure detection under long-term conditional variations. Furthermore, we evaluate six representative local feature matching methods (handcrafted and learning-based) under the benchmark, with in-depth analysis for limitations and future directions.
Authors: Mustafa Doga Dogan
Abstract: The convergence of the physical and digital realms has ushered in a new era of immersive experiences and seamless interactions. As the boundaries between the real world and virtual environments blur and result in a "mixed reality," there arises a need for robust and efficient methods to connect physical objects with their virtual counterparts. In this thesis, we present a novel approach to bridging this gap through the design, fabrication, and detection of embedded machine-readable markers. We categorize the proposed marking approaches into three distinct categories: natural markers, structural markers, and internal markers. Natural markers, such as those used in SensiCut, are inherent fingerprints of objects repurposed as machine-readable identifiers, while structural markers, such as StructCode and G-ID, leverage the structural artifacts in objects that emerge during the fabrication process itself. Internal markers, such as InfraredTag and BrightMarker, are embedded inside fabricated objects using specialized materials. Leveraging a combination of methods from computer vision, machine learning, computational imaging, and material science, the presented approaches offer robust and versatile solutions for object identification, tracking, and interaction. These markers, seamlessly integrated into real-world objects, effectively communicate an object's identity, origin, function, and interaction, functioning as gateways to "ubiquitous metadata" - a concept where metadata is embedded into physical objects, similar to metadata in digital files. Across the different chapters, we demonstrate the applications of the presented methods in diverse domains, including product design, manufacturing, retail, logistics, education, entertainment, security, and sustainability.
Authors: Luigi Capogrosso, Enrico Fraccaroli, Giulio Petrozziello, Francesco Setti, Samarjit Chakraborty, Franco Fummi, Marco Cristani
Abstract: In the past decade, Deep Neural Networks (DNNs) achieved state-of-the-art performance in a broad range of problems, spanning from object classification and action recognition to smart building and healthcare. The flexibility that makes DNNs such a pervasive technology comes at a price: the computational requirements preclude their deployment on most of the resource-constrained edge devices available today to solve real-time and real-world tasks. This paper introduces a novel approach to address this challenge by combining the concept of predefined sparsity with Split Computing (SC) and Early Exit (EE). In particular, SC aims at splitting a DNN with a part of it deployed on an edge device and the rest on a remote server. Instead, EE allows the system to stop using the remote server and rely solely on the edge device's computation if the answer is already good enough. Specifically, how to apply such a predefined sparsity to a SC and EE paradigm has never been studied. This paper studies this problem and shows how predefined sparsity significantly reduces the computational, storage, and energy burdens during the training and inference phases, regardless of the hardware platform. This makes it a valuable approach for enhancing the performance of SC and EE applications. Experimental results showcase reductions exceeding 4x in storage and computational complexity without compromising performance. The source code is available at https://github.com/intelligolabs/sparsity_sc_ee.
Authors: Daoyuan Chen, Haibin Wang, Yilun Huang, Ce Ge, Yaliang Li, Bolin Ding, Jingren Zhou
Abstract: The emergence of large-scale multi-modal generative models has drastically advanced artificial intelligence, introducing unprecedented levels of performance and functionality. However, optimizing these models remains challenging due to historically isolated paths of model-centric and data-centric developments, leading to suboptimal outcomes and inefficient resource utilization. In response, we present a novel sandbox suite tailored for integrated data-model co-development. This sandbox provides a comprehensive experimental platform, enabling rapid iteration and insight-driven refinement of both data and models. Our proposed "Probe-Analyze-Refine" workflow, validated through applications on state-of-the-art LLaVA-like and DiT based models, yields significant performance boosts, such as topping the VBench leaderboard. We also uncover fruitful insights gleaned from exhaustive benchmarks, shedding light on the critical interplay between data quality, diversity, and model behavior. With the hope of fostering deeper understanding and future progress in multi-modal data and generative modeling, our codes, datasets, and models are maintained and accessible at https://github.com/modelscope/data-juicer/blob/main/docs/Sandbox.md.
URLs: https://github.com/modelscope/data-juicer/blob/main/docs/Sandbox.md.
Authors: Biao Zhang, Garrett Tanzer, Orhan Firat
Abstract: Sign language translation (SLT) addresses the problem of translating information from a sign language in video to a spoken language in text. Existing studies, while showing progress, are often limited to narrow domains and/or few sign languages and struggle with open-domain tasks. In this paper, we push forward the frontier of SLT by scaling pretraining data, model size, and number of translation directions. We perform large-scale SLT pretraining on different data including 1) noisy multilingual YouTube SLT data, 2) parallel text corpora, and 3) SLT data augmented by translating video captions to other languages with off-the-shelf machine translation models. We unify different pretraining tasks with task-specific prompts under the encoder-decoder architecture, and initialize the SLT model with pretrained (m/By)T5 models across model sizes. SLT pretraining results on How2Sign and FLEURS-ASL#0 (ASL to 42 spoken languages) demonstrate the significance of data/model scaling and cross-lingual cross-modal transfer, as well as the feasibility of zero-shot SLT. We finetune the pretrained SLT models on 5 downstream open-domain SLT benchmarks covering 5 sign languages. Experiments show substantial quality improvements over the vanilla baselines, surpassing the previous state-of-the-art (SOTA) by wide margins.
Authors: Luca Tirel, Ali Mohamed Ali, Hashim A. Hashim
Abstract: This paper introduces a novel approach to image denoising that leverages the advantages of Generative Adversarial Networks (GANs). Specifically, we propose a model that combines elements of the Pix2Pix model and the Wasserstein GAN (WGAN) with Gradient Penalty (WGAN-GP). This hybrid framework seeks to capitalize on the denoising capabilities of conditional GANs, as demonstrated in the Pix2Pix model, while mitigating the need for an exhaustive search for optimal hyperparameters that could potentially ruin the stability of the learning process. In the proposed method, the GAN's generator is employed to produce denoised images, harnessing the power of a conditional GAN for noise reduction. Simultaneously, the implementation of the Lipschitz continuity constraint during updates, as featured in WGAN-GP, aids in reducing susceptibility to mode collapse. This innovative design allows the proposed model to benefit from the strong points of both Pix2Pix and WGAN-GP, generating superior denoising results while ensuring training stability. Drawing on previous work on image-to-image translation and GAN stabilization techniques, the proposed research highlights the potential of GANs as a general-purpose solution for denoising. The paper details the development and testing of this model, showcasing its effectiveness through numerical experiments. The dataset was created by adding synthetic noise to clean images. Numerical results based on real-world dataset validation underscore the efficacy of this approach in image-denoising tasks, exhibiting significant enhancements over traditional techniques. Notably, the proposed model demonstrates strong generalization capabilities, performing effectively even when trained with synthetic noise.
Authors: Jiafeng Wang, Zhaoyu Chen, Kaixun Jiang, Dingkang Yang, Lingyi Hong, Pinxue Guo, Haijing Guo, Wenqiang Zhang
Abstract: Deep Neural Networks (DNNs) are vulnerable to adversarial examples, which are crafted by adding human-imperceptible perturbations to the benign inputs. Simultaneously, adversarial examples exhibit transferability across models, enabling practical black-box attacks. However, existing methods are still incapable of achieving the desired transfer attack performance. In this work, focusing on gradient optimization and consistency, we analyse the gradient elimination phenomenon as well as the local momentum optimum dilemma. To tackle these challenges, we introduce Global Momentum Initialization (GI), providing global momentum knowledge to mitigate gradient elimination. Specifically, we perform gradient pre-convergence before the attack and a global search during this stage. GI seamlessly integrates with existing transfer methods, significantly improving the success rate of transfer attacks by an average of 6.4% under various advanced defense mechanisms compared to the state-of-the-art method. Ultimately, GI demonstrates strong transferability in both image and video attack domains. Particularly, when attacking advanced defense methods in the image domain, it achieves an average attack success rate of 95.4%. The code is available at $\href{https://github.com/Omenzychen/Global-Momentum-Initialization}{https://github.com/Omenzychen/Global-Momentum-Initialization}$.
URLs: https://github.com/Omenzychen/Global-Momentum-Initialization, https://github.com/Omenzychen/Global-Momentum-Initialization
Authors: Geonmo Gu, Sanghyuk Chun, Wonjae Kim, HeeJae Jun, Yoohoon Kang, Sangdoo Yun
Abstract: This paper proposes a novel diffusion-based model, CompoDiff, for solving zero-shot Composed Image Retrieval (ZS-CIR) with latent diffusion. This paper also introduces a new synthetic dataset, named SynthTriplets18M, with 18.8 million reference images, conditions, and corresponding target image triplets to train CIR models. CompoDiff and SynthTriplets18M tackle the shortages of the previous CIR approaches, such as poor generalizability due to the small dataset scale and the limited types of conditions. CompoDiff not only achieves a new state-of-the-art on four ZS-CIR benchmarks, including FashionIQ, CIRR, CIRCO, and GeneCIS, but also enables a more versatile and controllable CIR by accepting various conditions, such as negative text, and image mask conditions. CompoDiff also shows the controllability of the condition strength between text and image queries and the trade-off between inference speed and performance, which are unavailable with existing CIR methods. The code and dataset are available at https://github.com/navervision/CompoDiff
Authors: Francesco Croce, Naman D Singh, Matthias Hein
Abstract: Adversarial robustness has been studied extensively in image classification, especially for the $\ell_\infty$-threat model, but significantly less so for related tasks such as object detection and semantic segmentation, where attacks turn out to be a much harder optimization problem than for image classification. We propose several problem-specific novel attacks minimizing different metrics in accuracy and mIoU. The ensemble of our attacks, SEA, shows that existing attacks severely overestimate the robustness of semantic segmentation models. Surprisingly, existing attempts of adversarial training for semantic segmentation models turn out to be weak or even completely non-robust. We investigate why previous adaptations of adversarial training to semantic segmentation failed and show how recently proposed robust ImageNet backbones can be used to obtain adversarially robust semantic segmentation models with up to six times less training time for PASCAL-VOC and the more challenging ADE20k. The associated code and robust models are available at https://github.com/nmndeep/robust-segmentation
Authors: Spyros Gidaris, Andrei Bursuc, Oriane Simeoni, Antonin Vobecky, Nikos Komodakis, Matthieu Cord, Patrick P\'erez
Abstract: Self-supervised learning can be used for mitigating the greedy needs of Vision Transformer networks for very large fully-annotated datasets. Different classes of self-supervised learning offer representations with either good contextual reasoning properties, e.g., using masked image modeling strategies, or invariance to image perturbations, e.g., with contrastive methods. In this work, we propose a single-stage and standalone method, MOCA, which unifies both desired properties using novel mask-and-predict objectives defined with high-level features (instead of pixel-level details). Moreover, we show how to effectively employ both learning paradigms in a synergistic and computation-efficient way. Doing so, we achieve new state-of-the-art results on low-shot settings and strong experimental results in various evaluation protocols with a training that is at least 3 times faster than prior methods. We provide the implementation code at https://github.com/valeoai/MOCA.
Authors: Yiming Wu, Ruixiang Li, Zequn Qin, Xinhai Zhao, Xi Li
Abstract: Vision-based Bird's Eye View (BEV) representation is an emerging perception formulation for autonomous driving. The core challenge is to construct BEV space with multi-camera features, which is a one-to-many ill-posed problem. Diving into all previous BEV representation generation methods, we found that most of them fall into two types: modeling depths in image views or modeling heights in the BEV space, mostly in an implicit way. In this work, we propose to explicitly model heights in the BEV space, which needs no extra data like LiDAR and can fit arbitrary camera rigs and types compared to modeling depths. Theoretically, we give proof of the equivalence between height-based methods and depth-based methods. Considering the equivalence and some advantages of modeling heights, we propose HeightFormer, which models heights and uncertainties in a self-recursive way. Without any extra data, the proposed HeightFormer could estimate heights in BEV accurately. Benchmark results show that the performance of HeightFormer achieves SOTA compared with those camera-only methods.
Authors: Xinyu Lyu, Lianli Gao, Junlin Xie, Pengpeng Zeng, Yulu Tian, Jie Shao, Heng Tao Shen
Abstract: Existing Unbiased Scene Graph Generation (USGG) methods only focus on addressing the predicate-level imbalance that high-frequency classes dominate predictions of rare ones, while overlooking the concept-level imbalance. Actually, even if predicates themselves are balanced, there is still a significant concept-imbalance within them due to the long-tailed distribution of contexts (i.e., subject-object combinations). This concept-level imbalance poses a more pervasive and challenging issue compared to the predicate-level imbalance since subject-object pairs are inherently complex in combinations. Hence, we introduce a novel research problem: Generalized Unbiased Scene Graph Generation (G-USGG), which takes into account both predicate-level and concept-level imbalance. To the end, we propose the Multi-Concept Learning (MCL) framework, which ensures a balanced learning process across rare/ uncommon/ common concepts. MCL first quantifies the concept-level imbalance across predicates in terms of different amounts of concepts, representing as multiple concept-prototypes within the same class. It then effectively learns concept-prototypes by applying the Concept Regularization (CR) technique. Furthermore, to achieve balanced learning over different concepts, we introduce the Balanced Prototypical Memory (BPM), which guides SGG models to generate balanced representations for concept-prototypes. Extensive experiments demonstrate the remarkable efficacy of our model-agnostic strategy in enhancing the performance of benchmark models on both VG-SGG and OI-SGG datasets, leading to new state-of-the-art achievements in two key aspects: predicate-level unbiased relation recognition and concept-level compositional generability.
Authors: Ozan Unal, Christos Sakaridis, Suman Saha, Luc Van Gool
Abstract: 3D visual grounding is the task of localizing the object in a 3D scene which is referred by a description in natural language. With a wide range of applications ranging from autonomous indoor robotics to AR/VR, the task has recently risen in popularity. A common formulation to tackle 3D visual grounding is grounding-by-detection, where localization is done via bounding boxes. However, for real-life applications that require physical interactions, a bounding box insufficiently describes the geometry of an object. We therefore tackle the problem of dense 3D visual grounding, i.e. referral-based 3D instance segmentation. We propose a dense 3D grounding network ConcreteNet, featuring four novel stand-alone modules that aim to improve grounding performance for challenging repetitive instances, i.e. instances with distractors of the same semantic class. First, we introduce a bottom-up attentive fusion module that aims to disambiguate inter-instance relational cues, next, we construct a contrastive training scheme to induce separation in the latent space, we then resolve view-dependent utterances via a learned global camera token, and finally we employ multi-view ensembling to improve referred mask quality. ConcreteNet ranks 1st on the challenging ScanRefer online benchmark and has won the ICCV 3rd Workshop on Language for 3D Scenes "3D Object Localization" challenge. Our code is available at ouenal.github.io/concretenet/.
Authors: Xiangyu Chen, Zheyuan Li, Yuandong Pu, Yihao Liu, Jiantao Zhou, Yu Qiao, Chao Dong
Abstract: Despite the significant progress made by deep models in various image restoration tasks, existing image restoration networks still face challenges in terms of task generality. An intuitive manifestation is that networks which excel in certain tasks often fail to deliver satisfactory results in others. To illustrate this point, we select five representative networks and conduct a comparative study on five classic image restoration tasks. First, we provide a detailed explanation of the characteristics of different image restoration tasks and backbone networks. Following this, we present the benchmark results and analyze the reasons behind the performance disparity of different models across various tasks. Drawing from this comparative study, we propose that a general image restoration backbone network needs to meet the functional requirements of diverse tasks. Based on this principle, we design a new general image restoration backbone network, X-Restormer. Extensive experiments demonstrate that X-Restormer possesses good task generality and achieves state-of-the-art performance across a variety of tasks.
Authors: Camillo Quattrocchi, Antonino Furnari, Daniele Di Mauro, Mario Valerio Giuffrida, Giovanni Maria Farinella
Abstract: We consider the problem of transferring a temporal action segmentation system initially designed for exocentric (fixed) cameras to an egocentric scenario, where wearable cameras capture video data. The conventional supervised approach requires the collection and labeling of a new set of egocentric videos to adapt the model, which is costly and time-consuming. Instead, we propose a novel methodology which performs the adaptation leveraging existing labeled exocentric videos and a new set of unlabeled, synchronized exocentric-egocentric video pairs, for which temporal action segmentation annotations do not need to be collected. We implement the proposed methodology with an approach based on knowledge distillation, which we investigate both at the feature and Temporal Action Segmentation model level. Experiments on Assembly101 and EgoExo4D demonstrate the effectiveness of the proposed method against classic unsupervised domain adaptation and temporal alignment approaches. Without bells and whistles, our best model performs on par with supervised approaches trained on labeled egocentric data, without ever seeing a single egocentric label, achieving a +15.99 improvement in the edit score (28.59 vs 12.60) on the Assembly101 dataset compared to a baseline model trained solely on exocentric data. In similar settings, our method also improves edit score by +3.32 on the challenging EgoExo4D benchmark. Code is available here: https://github.com/fpv-iplab/synchronization-is-all-you-need.
URLs: https://github.com/fpv-iplab/synchronization-is-all-you-need.
Authors: Rosario Leonardi, Antonino Furnari, Francesco Ragusa, Giovanni Maria Farinella
Abstract: In this study, we investigate the effectiveness of synthetic data in enhancing egocentric hand-object interaction detection. Via extensive experiments and comparative analyses on three egocentric datasets, VISOR, EgoHOS, and ENIGMA-51, our findings reveal how to exploit synthetic data for the HOI detection task when real labeled data are scarce or unavailable. Specifically, by leveraging only 10% of real labeled data, we achieve improvements in Overall AP compared to baselines trained exclusively on real data of: +5.67% on EPIC-KITCHENS VISOR, +8.24% on EgoHOS, and +11.69% on ENIGMA-51. Our analysis is supported by a novel data generation pipeline and the newly introduced HOI-Synth benchmark which augments existing datasets with synthetic images of hand-object interactions automatically labeled with hand-object contact states, bounding boxes, and pixel-wise segmentation masks. Data, code, and data generation tools to support future research are released at: https://fpv-iplab.github.io/HOI-Synth/.
Authors: Zhouxia Wang, Ziyang Yuan, Xintao Wang, Tianshui Chen, Menghan Xia, Ping Luo, Ying Shan
Abstract: Motions in a video primarily consist of camera motion, induced by camera movement, and object motion, resulting from object movement. Accurate control of both camera and object motion is essential for video generation. However, existing works either mainly focus on one type of motion or do not clearly distinguish between the two, limiting their control capabilities and diversity. Therefore, this paper presents MotionCtrl, a unified and flexible motion controller for video generation designed to effectively and independently control camera and object motion. The architecture and training strategy of MotionCtrl are carefully devised, taking into account the inherent properties of camera motion, object motion, and imperfect training data. Compared to previous methods, MotionCtrl offers three main advantages: 1) It effectively and independently controls camera motion and object motion, enabling more fine-grained motion control and facilitating flexible and diverse combinations of both types of motion. 2) Its motion conditions are determined by camera poses and trajectories, which are appearance-free and minimally impact the appearance or shape of objects in generated videos. 3) It is a relatively generalizable model that can adapt to a wide array of camera poses and trajectories once trained. Extensive qualitative and quantitative experiments have been conducted to demonstrate the superiority of MotionCtrl over existing methods. Project Page: https://wzhouxiff.github.io/projects/MotionCtrl/
Authors: Gyeongrok Oh, Jaehwan Jeong, Sieun Kim, Wonmin Byeon, Jinkyu Kim, Sungwoong Kim, Sangpil Kim
Abstract: We introduce a novel diffusion-based video generation method, generating a video showing multiple events given multiple individual sentences from the user. Our method does not require a large-scale video dataset since our method uses a pre-trained diffusion-based text-to-video generative model without a fine-tuning process. Specifically, we propose a last frame-aware diffusion process to preserve visual coherence between consecutive videos where each video consists of different events by initializing the latent and simultaneously adjusting noise in the latent to enhance the motion dynamic in a generated video. Furthermore, we find that the iterative update of latent vectors by referring to all the preceding frames maintains the global appearance across the frames in a video clip. To handle dynamic text input for video generation, we utilize a novel prompt generator that transfers course text messages from the user into the multiple optimal prompts for the text-to-video diffusion model. Extensive experiments and user studies show that our proposed method is superior to other video-generative models in terms of temporal coherency of content and semantics. Video examples are available on our project page: https://kuai-lab.github.io/eccv2024mevg.
Authors: Zhiyu Pan, Zhicheng Zhong, Wenxuan Guo, Yifan Chen, Jianjiang Feng, Jie Zhou
Abstract: Several methods have been proposed to estimate 3D human pose from multi-view images, achieving satisfactory performance on public datasets collected under relatively simple conditions. However, there are limited approaches studying extracting 3D human skeletons from multimodal inputs, such as RGB and point cloud data. To address this gap, we introduce LiCamPose, a pipeline that integrates multi-view RGB and sparse point cloud information to estimate robust 3D human poses via single frame. We demonstrate the effectiveness of the volumetric architecture in combining these modalities. Furthermore, to circumvent the need for manually labeled 3D human pose annotations, we develop a synthetic dataset generator for pretraining and design an unsupervised domain adaptation strategy to train a 3D human pose estimator without manual annotations. To validate the generalization capability of our method, LiCamPose is evaluated on four datasets, including two public datasets, one synthetic dataset, and one challenging self-collected dataset named BasketBall, covering diverse scenarios. The results demonstrate that LiCamPose exhibits great generalization performance and significant application potential. The code, generator, and datasets will be made available upon acceptance of this paper.
Authors: Chenyang Qi, Zhengzhong Tu, Keren Ye, Mauricio Delbracio, Peyman Milanfar, Qifeng Chen, Hossein Talebi
Abstract: Text-driven diffusion models have become increasingly popular for various image editing tasks, including inpainting, stylization, and object replacement. However, it still remains an open research problem to adopt this language-vision paradigm for more fine-level image processing tasks, such as denoising, super-resolution, deblurring, and compression artifact removal. In this paper, we develop SPIRE, a Semantic and restoration Prompt-driven Image Restoration framework that leverages natural language as a user-friendly interface to control the image restoration process. We consider the capacity of prompt information in two dimensions. First, we use content-related prompts to enhance the semantic alignment, effectively alleviating identity ambiguity in the restoration outcomes. Second, our approach is the first framework that supports fine-level instruction through language-based quantitative specification of the restoration strength, without the need for explicit task-specific design. In addition, we introduce a novel fusion mechanism that augments the existing ControlNet architecture by learning to rescale the generative prior, thereby achieving better restoration fidelity. Our extensive experiments demonstrate the superior restoration performance of SPIRE compared to the state of the arts, alongside offering the flexibility of text-based control over the restoration effects.
Authors: Jaeyeul Kim, Jungwan Woo, Jeonghoon Kim, Sunghoon Im
Abstract: In the realm of LiDAR-based perception, significant strides have been made, yet domain generalization remains a substantial challenge. The performance often deteriorates when models are applied to unfamiliar datasets with different LiDAR sensors or deployed in new environments, primarily due to variations in point cloud density distributions. To tackle this challenge, we propose a Density Discriminative Feature Embedding (DDFE) module, capitalizing on the observation that a single source LiDAR point cloud encompasses a spectrum of densities. The DDFE module is meticulously designed to extract density-specific features within a single source domain, facilitating the recognition of objects sharing similar density characteristics across different LiDAR sensors. In addition, we introduce a simple yet effective density augmentation technique aimed at expanding the spectrum of density in source data, thereby enhancing the capabilities of the DDFE. Our DDFE stands out as a versatile and lightweight domain generalization module. It can be seamlessly integrated into various 3D backbone networks, where it has demonstrated superior performance over current state-of-the-art domain generalization methods. Code is available at https://github.com/dgist-cvlab/MultiDensityDG.
Authors: Jens Naumann, Binbin Xu, Stefan Leutenegger, Xingxing Zuo
Abstract: We introduce a novel monocular visual odometry (VO) system, NeRF-VO, that integrates learning-based sparse visual odometry for low-latency camera tracking and a neural radiance scene representation for fine-detailed dense reconstruction and novel view synthesis. Our system initializes camera poses using sparse visual odometry and obtains view-dependent dense geometry priors from a monocular prediction network. We harmonize the scale of poses and dense geometry, treating them as supervisory cues to train a neural implicit scene representation. NeRF-VO demonstrates exceptional performance in both photometric and geometric fidelity of the scene representation by jointly optimizing a sliding window of keyframed poses and the underlying dense geometry, which is accomplished through training the radiance field with volume rendering. We surpass SOTA methods in pose estimation accuracy, novel view synthesis fidelity, and dense reconstruction quality across a variety of synthetic and real-world datasets while achieving a higher camera tracking frequency and consuming less GPU memory.
Authors: Yunzhi Yan, Haotong Lin, Chenxu Zhou, Weijie Wang, Haiyang Sun, Kun Zhan, Xianpeng Lang, Xiaowei Zhou, Sida Peng
Abstract: This paper aims to tackle the problem of modeling dynamic urban streets for autonomous driving scenes. Recent methods extend NeRF by incorporating tracked vehicle poses to animate vehicles, enabling photo-realistic view synthesis of dynamic urban street scenes. However, significant limitations are their slow training and rendering speed. We introduce Street Gaussians, a new explicit scene representation that tackles these limitations. Specifically, the dynamic urban scene is represented as a set of point clouds equipped with semantic logits and 3D Gaussians, each associated with either a foreground vehicle or the background. To model the dynamics of foreground object vehicles, each object point cloud is optimized with optimizable tracked poses, along with a 4D spherical harmonics model for the dynamic appearance. The explicit representation allows easy composition of object vehicles and background, which in turn allows for scene editing operations and rendering at 135 FPS (1066 $\times$ 1600 resolution) within half an hour of training. The proposed method is evaluated on multiple challenging benchmarks, including KITTI and Waymo Open datasets. Experiments show that the proposed method consistently outperforms state-of-the-art methods across all datasets. The code will be released to ensure reproducibility.
Authors: Namitha Padmanabhan, Matthew Gwilliam, Pulkit Kumar, Shishira R Maiya, Max Ehrlich, Abhinav Shrivastava
Abstract: The many variations of Implicit Neural Representations (INRs), where a neural network is trained as a continuous representation of a signal, have tremendous practical utility for downstream tasks including novel view synthesis, video compression, and image super-resolution. Unfortunately, the inner workings of these networks are seriously under-studied. Our work, eXplaining the Implicit Neural Canvas (XINC), is a unified framework for explaining properties of INRs by examining the strength of each neuron's contribution to each output pixel. We call the aggregate of these contribution maps the Implicit Neural Canvas and we use this concept to demonstrate that the INRs we study learn to "see" the frames they represent in surprising ways. For example, INRs tend to have highly distributed representations. While lacking high-level object semantics, they have a significant bias for color and edges, and are almost entirely space-agnostic. We arrive at our conclusions by examining how objects are represented across time in video INRs, using clustering to visualize similar neurons across layers and architectures, and show that this is dominated by motion. These insights demonstrate the general usefulness of our analysis framework. Our project page is available at https://namithap10.github.io/xinc.
Authors: Wouter Van Gansbeke, Bert De Brabandere
Abstract: Panoptic and instance segmentation networks are often trained with specialized object detection modules, complex loss functions, and ad-hoc post-processing steps to manage the permutation-invariance of the instance masks. This work builds upon Stable Diffusion and proposes a latent diffusion approach for panoptic segmentation, resulting in a simple architecture that omits these complexities. Our training consists of two steps: (1) training a shallow autoencoder to project the segmentation masks to latent space; (2) training a diffusion model to allow image-conditioned sampling in latent space. This generative approach unlocks the exploration of mask completion or inpainting. The experimental validation on COCO and ADE20k yields strong segmentation results. Finally, we demonstrate our model's adaptability to multi-tasking by introducing learnable task embeddings.
Authors: Rohan Wadhawan, Hritik Bansal, Kai-Wei Chang, Nanyun Peng
Abstract: Many real-world tasks require an agent to reason jointly over text and visual objects, (e.g., navigating in public spaces), which we refer to as context-sensitive text-rich visual reasoning. Specifically, these tasks require an understanding of the context in which the text interacts with visual elements within an image. However, there is a lack of existing datasets to benchmark the state-of-the-art multimodal models' capability on context-sensitive text-rich visual reasoning. In this paper, we introduce ConTextual, a novel dataset featuring human-crafted instructions that require context-sensitive reasoning for text-rich images. We conduct experiments to assess the performance of 14 foundation models (GPT-4V, Gemini-Pro-Vision, LLaVA-Next) and establish a human performance baseline. Further, we perform human evaluations of the model responses and observe a significant performance gap of 30.8% between GPT-4V (the current best-performing Large Multimodal Model) and human performance. Our fine-grained analysis reveals that GPT-4V encounters difficulties interpreting time-related data and infographics. However, it demonstrates proficiency in comprehending abstract visual contexts such as memes and quotes. Finally, our qualitative analysis uncovers various factors contributing to poor performance including lack of precise visual perception and hallucinations. Our dataset, code, and leaderboard can be found on the project page https://con-textual.github.io/
Authors: Dilxat Muhtar, Zhenshi Li, Feng Gu, Xueliang Zhang, Pengfeng Xiao
Abstract: The revolutionary capabilities of large language models (LLMs) have paved the way for multimodal large language models (MLLMs) and fostered diverse applications across various specialized domains. In the remote sensing (RS) field, however, the diverse geographical landscapes and varied objects in RS imagery are not adequately considered in recent MLLM endeavors. To bridge this gap, we construct a large-scale RS image-text dataset, LHRS-Align, and an informative RS-specific instruction dataset, LHRS-Instruct, leveraging the extensive volunteered geographic information (VGI) and globally available RS images. Building on this foundation, we introduce LHRS-Bot, an MLLM tailored for RS image understanding through a novel multi-level vision-language alignment strategy and a curriculum learning method. Additionally, we introduce LHRS-Bench, a benchmark for thoroughly evaluating MLLMs' abilities in RS image understanding. Comprehensive experiments demonstrate that LHRS-Bot exhibits a profound understanding of RS images and the ability to perform nuanced reasoning within the RS domain.
Authors: Yuqian Fu, Yu Wang, Yixuan Pan, Lian Huai, Xingyu Qiu, Zeyu Shangguan, Tong Liu, Yanwei Fu, Luc Van Gool, Xingqun Jiang
Abstract: This paper studies the challenging cross-domain few-shot object detection (CD-FSOD), aiming to develop an accurate object detector for novel domains with minimal labeled examples. While transformer-based open-set detectors, such as DE-ViT, show promise in traditional few-shot object detection, their generalization to CD-FSOD remains unclear: 1) can such open-set detection methods easily generalize to CD-FSOD? 2) If not, how can models be enhanced when facing huge domain gaps? To answer the first question, we employ measures including style, inter-class variance (ICV), and indefinable boundaries (IB) to understand the domain gap. Based on these measures, we establish a new benchmark named CD-FSOD to evaluate object detection methods, revealing that most of the current approaches fail to generalize across domains. Technically, we observe that the performance decline is associated with our proposed measures: style, ICV, and IB. Consequently, we propose several novel modules to address these issues. First, the learnable instance features align initial fixed instances with target categories, enhancing feature distinctiveness. Second, the instance reweighting module assigns higher importance to high-quality instances with slight IB. Third, the domain prompter encourages features resilient to different styles by synthesizing imaginary domains without altering semantic contents. These techniques collectively contribute to the development of the Cross-Domain Vision Transformer for CD-FSOD (CD-ViTO), significantly improving upon the base DE-ViT. Experimental results validate the efficacy of our model.
Authors: Aoran Xiao, Weihao Xuan, Heli Qi, Yun Xing, Ruijie Ren, Xiaoqin Zhang, Ling Shao, Shijian Lu
Abstract: The recent Segment Anything Model (SAM) has demonstrated remarkable zero-shot capability and flexible geometric prompting in general image segmentation. However, SAM often struggles when handling various unconventional images, such as aerial, medical, and non-RGB images. This paper presents CAT-SAM, a ConditionAl Tuning network that adapts SAM toward various unconventional target tasks with just few-shot target samples. CAT-SAM freezes the entire SAM and adapts its mask decoder and image encoder simultaneously with a small number of learnable parameters. The core design is a prompt bridge structure that enables decoder-conditioned joint tuning of the heavyweight image encoder and the lightweight mask decoder. The bridging maps the prompt token of the mask decoder to the image encoder, fostering synergic adaptation of the encoder and the decoder with mutual benefits. We develop two representative tuning strategies for the image encoder which leads to two CAT-SAM variants: one injecting learnable prompt tokens in the input space and the other inserting lightweight adapter networks. Extensive experiments over 11 unconventional tasks show that both CAT-SAM variants achieve superior target segmentation performance consistently even under the very challenging one-shot adaptation setup. Project page: https://xiaoaoran.github.io/projects/CAT-SAM
Authors: Shikun Ban, Juling Fan, Xiaoxuan Ma, Wentao Zhu, Yu Qiao, Yizhou Wang
Abstract: Estimating robot pose from RGB images is a crucial problem in computer vision and robotics. While previous methods have achieved promising performance, most of them presume full knowledge of robot internal states, e.g. ground-truth robot joint angles. However, this assumption is not always valid in practical situations. In real-world applications such as multi-robot collaboration or human-robot interaction, the robot joint states might not be shared or could be unreliable. On the other hand, existing approaches that estimate robot pose without joint state priors suffer from heavy computation burdens and thus cannot support real-time applications. This work introduces an efficient framework for real-time robot pose estimation from RGB images without requiring known robot states. Our method estimates camera-to-robot rotation, robot state parameters, keypoint locations, and root depth, employing a neural network module for each task to facilitate learning and sim-to-real transfer. Notably, it achieves inference in a single feed-forward pass without iterative optimization. Our approach offers a 12-time speed increase with state-of-the-art accuracy, enabling real-time holistic robot pose estimation for the first time. Code and models are available at https://github.com/Oliverbansk/Holistic-Robot-Pose-Estimation.
URLs: https://github.com/Oliverbansk/Holistic-Robot-Pose-Estimation.
Authors: Yang Qian, Yinan Sun, Ali Kargarandehkordi, Parnian Azizian, Onur Cezmi Mutlu, Saimourya Surabhi, Pingyi Chen, Zain Jabbar, Dennis Paul Wall, Peter Washington
Abstract: The increasing variety and quantity of tagged multimedia content on a variety of online platforms offer a unique opportunity to advance the field of human action recognition. In this study, we utilize 283,582 unique, unlabeled TikTok video clips, categorized into 386 hashtags, to train a domain-specific foundation model for action recognition. We employ VideoMAE V2, an advanced model integrating Masked Autoencoders (MAE) with Vision Transformers (ViT), pre-trained on this diverse collection of unstructured videos. Our model, fine-tuned on established action recognition benchmarks such as UCF101 and HMDB51, achieves state-of-the-art results: 99.05% on UCF101, 86.08% on HMDB51, 85.51% on Kinetics-400, and 74.27% on Something-Something V2 using the ViT-giant backbone. These results highlight the potential of using unstructured and unlabeled videos as a valuable source of diverse and dynamic content for training foundation models. Our investigation confirms that while initial increases in pre-training data volume significantly enhance model performance, the gains diminish as the dataset size continues to expand. Our findings emphasize two critical axioms in self-supervised learning for computer vision: (1) additional pre-training data can yield diminishing benefits for some datasets and (2) quality is more important than quantity in self-supervised learning, especially when building foundation models.
Authors: Yizhe Xiong, Hui Chen, Tianxiang Hao, Zijia Lin, Jungong Han, Yuesong Zhang, Guoxin Wang, Yongjun Bao, Guiguang Ding
Abstract: Recently, the scale of transformers has grown rapidly, which introduces considerable challenges in terms of training overhead and inference efficiency in the scope of task adaptation. Existing works, namely Parameter-Efficient Fine-Tuning (PEFT) and model compression, have separately investigated the challenges. However, PEFT cannot guarantee the inference efficiency of the original backbone, especially for large-scale models. Model compression requires significant training costs for structure searching and re-training. Consequently, a simple combination of them cannot guarantee accomplishing both training efficiency and inference efficiency with minimal costs. In this paper, we propose a novel Parallel Yielding Re-Activation (PYRA) method for such a challenge of training-inference efficient task adaptation. PYRA first utilizes parallel yielding adaptive weights to comprehensively perceive the data distribution in downstream tasks. A re-activation strategy for token modulation is then applied for tokens to be merged, leading to calibrated token features. Extensive experiments demonstrate that PYRA outperforms all competing methods under both low compression rate and high compression rate, demonstrating its effectiveness and superiority in maintaining both training efficiency and inference efficiency for large-scale foundation models. Our code is available at https://github.com/THU-MIG/PYRA.
Authors: Huan-ang Gao, Mingju Gao, Jiaju Li, Wenyi Li, Rong Zhi, Hao Tang, Hao Zhao
Abstract: Semantic image synthesis (SIS) shows good promises for sensor simulation. However, current best practices in this field, based on GANs, have not yet reached the desired level of quality. As latent diffusion models make significant strides in image generation, we are prompted to evaluate ControlNet, a notable method for its dense control capabilities. Our investigation uncovered two primary issues with its results: the presence of weird sub-structures within large semantic areas and the misalignment of content with the semantic mask. Through empirical study, we pinpointed the cause of these problems as a mismatch between the noised training data distribution and the standard normal prior applied at the inference stage. To address this challenge, we developed specific noise priors for SIS, encompassing spatial, categorical, and a novel spatial-categorical joint prior for inference. This approach, which we have named SCP-Diff, has set new state-of-the-art results in SIS on Cityscapes, ADE20K and COCO-Stuff, yielding a FID as low as 10.53 on Cityscapes. The code and models can be accessed via the project page.
Authors: Nicolas Chahine, Sira Ferradans, Jean Ponce
Abstract: Blind image quality assessment (BIQA) approaches, while promising for automating image quality evaluation, often fall short in real-world scenarios due to their reliance on a generic quality standard applied uniformly across diverse images. This one-size-fits-all approach overlooks the crucial perceptual relationship between image content and quality, leading to a 'domain shift' challenge where a single quality metric inadequately represents various content types. Furthermore, BIQA techniques typically overlook the inherent differences in the human visual system among different observers. In response to these challenges, this paper introduces PICNIQ, a pairwise comparison framework designed to bypass the limitations of conventional BIQA by emphasizing relative, rather than absolute, quality assessment. PICNIQ is specifically designed to estimate the preference likelihood of quality between image pairs. By employing psychometric scaling algorithms, PICNIQ transforms pairwise comparisons into just-objectionable-difference (JOD) quality scores, offering a granular and interpretable measure of image quality. The proposed framework implements a deep learning architecture in combination with a specialized loss function, and a training strategy optimized for sparse pairwise comparison settings. We conduct our research using comparison matrices from the PIQ23 dataset, which are published in this paper. Our extensive experimental analysis showcases PICNIQ's broad applicability and competitive performance, highlighting its potential to set new standards in the field of BIQA.
Authors: Sarah Antiles, Sachin S. Talathi
Abstract: We present the Open Stamped Parts Dataset (OSPD), featuring synthetic and real images of stamped metal sheets for auto manufacturing. The real part images, captured from 7 cameras, consist of 7,980 unlabeled images and 1,680 labeled images. In addition, we have compiled a defect dataset by overlaying synthetically generated masks on 10\% of the holes. The synthetic dataset replicates the real manufacturing environment in terms of lighting and part placement relative to the cameras. The synthetic data includes 7,980 training images, 1,680 validation images and 1,680 test images, each with bounding box and segmentation mask annotations around all holes. 10\% of the holes in the synthetic data mimic defects generated in the real image dataset. We trained a hole-detection model on the synthetic-OSPD, achieving a modified recall score of 67.2\% and a precision of 94.4\% . We anticipate researchers in auto manufacturing use OSPD to advance the state of the art in defect detection of stamped holes in the metal-sheet stamping process. The dataset is available for download at: https://tinyurl.com/hm6xatd7.
Authors: Minkyu Choi, Harsh Goel, Mohammad Omama, Yunhao Yang, Sahil Shah, Sandeep Chinchali
Abstract: The unprecedented surge in video data production in recent years necessitates efficient tools to extract meaningful frames from videos for downstream tasks. Long-term temporal reasoning is a key desideratum for frame retrieval systems. While state-of-the-art foundation models, like VideoLLaMA and ViCLIP, are proficient in short-term semantic understanding, they surprisingly fail at long-term reasoning across frames. A key reason for this failure is that they intertwine per-frame perception and temporal reasoning into a single deep network. Hence, decoupling but co-designing semantic understanding and temporal reasoning is essential for efficient scene identification. We propose a system that leverages vision-language models for semantic understanding of individual frames but effectively reasons about the long-term evolution of events using state machines and temporal logic (TL) formulae that inherently capture memory. Our TL-based reasoning improves the F1 score of complex event identification by 9-15% compared to benchmarks that use GPT4 for reasoning on state-of-the-art self-driving datasets such as Waymo and NuScenes.
Authors: Liren He, Zhengkai Jiang, Jinlong Peng, Liang Liu, Qiangang Du, Xiaobin Hu, Wenbing Zhu, Mingmin Chi, Yabiao Wang, Chengjie Wang
Abstract: In the field of multi-class anomaly detection, reconstruction-based methods derived from single-class anomaly detection face the well-known challenge of "learning shortcuts", wherein the model fails to learn the patterns of normal samples as it should, opting instead for shortcuts such as identity mapping or artificial noise elimination. Consequently, the model becomes unable to reconstruct genuine anomalies as normal instances, resulting in a failure of anomaly detection. To counter this issue, we present a novel unified feature reconstruction-based anomaly detection framework termed RLR (Reconstruct features from a Learnable Reference representation). Unlike previous methods, RLR utilizes learnable reference representations to compel the model to learn normal feature patterns explicitly, thereby prevents the model from succumbing to the "learning shortcuts" issue. Additionally, RLR incorporates locality constraints into the learnable reference to facilitate more effective normal pattern capture and utilizes a masked learnable key attention mechanism to enhance robustness. Evaluation of RLR on the 15-category MVTec-AD dataset and the 12-category VisA dataset shows superior performance compared to state-of-the-art methods under the unified setting. The code of RLR will be publicly available.
Authors: Xianglong He, Junyi Chen, Sida Peng, Di Huang, Yangguang Li, Xiaoshui Huang, Chun Yuan, Wanli Ouyang, Tong He
Abstract: In recent years, 3D Gaussian splatting has emerged as a powerful technique for 3D reconstruction and generation, known for its fast and high-quality rendering capabilities. To address these shortcomings, this paper introduces a novel diffusion-based framework, GVGEN, designed to efficiently generate 3D Gaussian representations from text input. We propose two innovative techniques:(1) Structured Volumetric Representation. We first arrange disorganized 3D Gaussian points as a structured form GaussianVolume. This transformation allows the capture of intricate texture details within a volume composed of a fixed number of Gaussians. To better optimize the representation of these details, we propose a unique pruning and densifying method named the Candidate Pool Strategy, enhancing detail fidelity through selective optimization. (2) Coarse-to-fine Generation Pipeline. To simplify the generation of GaussianVolume and empower the model to generate instances with detailed 3D geometry, we propose a coarse-to-fine pipeline. It initially constructs a basic geometric structure, followed by the prediction of complete Gaussian attributes. Our framework, GVGEN, demonstrates superior performance in qualitative and quantitative assessments compared to existing 3D generation methods. Simultaneously, it maintains a fast generation speed ($\sim$7 seconds), effectively striking a balance between quality and efficiency. Our project page is: https://gvgen.github.io/
URLs: https://gvgen.github.io/
Authors: Byeongho Heo, Song Park, Dongyoon Han, Sangdoo Yun
Abstract: Rotary Position Embedding (RoPE) performs remarkably on language models, especially for length extrapolation of Transformers. However, the impacts of RoPE on computer vision domains have been underexplored, even though RoPE appears capable of enhancing Vision Transformer (ViT) performance in a way similar to the language domain. This study provides a comprehensive analysis of RoPE when applied to ViTs, utilizing practical implementations of RoPE for 2D vision data. The analysis reveals that RoPE demonstrates impressive extrapolation performance, i.e., maintaining precision while increasing image resolution at inference. It eventually leads to performance improvement for ImageNet-1k, COCO detection, and ADE-20k segmentation. We believe this study provides thorough guidelines to apply RoPE into ViT, promising improved backbone performance with minimal extra computational overhead. Our code and pre-trained models are available at https://github.com/naver-ai/rope-vit
Authors: Qiaoqiao Jin, Xuanhong Chen, Meiguang Jin, Ying Chen, Rui Shi, Yucheng Zheng, Yupeng Zhu, Bingbing Ni
Abstract: Contemporary makeup approaches primarily hinge on unpaired learning paradigms, yet they grapple with the challenges of inaccurate supervision (e.g., face misalignment) and sophisticated facial prompts (including face parsing, and landmark detection). These challenges prohibit low-cost deployment of facial makeup models, especially on mobile devices. To solve above problems, we propose a brand-new learning paradigm, termed "Data Amplify Learning (DAL)," alongside a compact makeup model named "TinyBeauty." The core idea of DAL lies in employing a Diffusion-based Data Amplifier (DDA) to "amplify" limited images for the model training, thereby enabling accurate pixel-to-pixel supervision with merely a handful of annotations. Two pivotal innovations in DDA facilitate the above training approach: (1) A Residual Diffusion Model (RDM) is designed to generate high-fidelity detail and circumvent the detail vanishing problem in the vanilla diffusion models; (2) A Fine-Grained Makeup Module (FGMM) is proposed to achieve precise makeup control and combination while retaining face identity. Coupled with DAL, TinyBeauty necessitates merely 80K parameters to achieve a state-of-the-art performance without intricate face prompts. Meanwhile, TinyBeauty achieves a remarkable inference speed of up to 460 fps on the iPhone 13. Extensive experiments show that DAL can produce highly competitive makeup models using only 5 image pairs.
Authors: Sisi Dai, Wenhao Li, Haowen Sun, Haibin Huang, Chongyang Ma, Hui Huang, Kai Xu, Ruizhen Hu
Abstract: In this study, we tackle the complex task of generating 3D human-object interactions (HOI) from textual descriptions in a zero-shot text-to-3D manner. We identify and address two key challenges: the unsatisfactory outcomes of direct text-to-3D methods in HOI, largely due to the lack of paired text-interaction data, and the inherent difficulties in simultaneously generating multiple concepts with complex spatial relationships. To effectively address these issues, we present InterFusion, a two-stage framework specifically designed for HOI generation. InterFusion involves human pose estimations derived from text as geometric priors, which simplifies the text-to-3D conversion process and introduces additional constraints for accurate object generation. At the first stage, InterFusion extracts 3D human poses from a synthesized image dataset depicting a wide range of interactions, subsequently mapping these poses to interaction descriptions. The second stage of InterFusion capitalizes on the latest developments in text-to-3D generation, enabling the production of realistic and high-quality 3D HOI scenes. This is achieved through a local-global optimization process, where the generation of human body and object is optimized separately, and jointly refined with a global optimization of the entire scene, ensuring a seamless and contextually coherent integration. Our experimental results affirm that InterFusion significantly outperforms existing state-of-the-art methods in 3D HOI generation.
Authors: Akshay Paruchuri, Samuel Ehrenstein, Shuxian Wang, Inbar Fried, Stephen M. Pizer, Marc Niethammer, Roni Sengupta
Abstract: Monocular depth estimation in endoscopy videos can enable assistive and robotic surgery to obtain better coverage of the organ and detection of various health issues. Despite promising progress on mainstream, natural image depth estimation, techniques perform poorly on endoscopy images due to a lack of strong geometric features and challenging illumination effects. In this paper, we utilize the photometric cues, i.e., the light emitted from an endoscope and reflected by the surface, to improve monocular depth estimation. We first create two novel loss functions with supervised and self-supervised variants that utilize a per-pixel shading representation. We then propose a novel depth refinement network (PPSNet) that leverages the same per-pixel shading representation. Finally, we introduce teacher-student transfer learning to produce better depth maps from both synthetic data with supervision and clinical data with self-supervision. We achieve state-of-the-art results on the C3VD dataset while estimating high-quality depth maps from clinical data. Our code, pre-trained models, and supplementary materials can be found on our project page: https://ppsnet.github.io/
Authors: Yutong Chen, Yifan Zhan, Zhihang Zhong, Wei Wang, Xiao Sun, Yu Qiao, Yinqiang Zheng
Abstract: Neural rendering techniques have significantly advanced 3D human body modeling. However, previous approaches often overlook dynamics induced by factors such as motion inertia, leading to challenges in scenarios like abrupt stops after rotation, where the pose remains static while the appearance changes. This limitation arises from reliance on a single pose as conditional input, resulting in ambiguity in mapping one pose to multiple appearances. In this study, we elucidate that variations in human appearance depend not only on the current frame's pose condition but also on past pose states. Therefore, we introduce Dyco, a novel method utilizing the delta pose sequence representation for non-rigid deformations and canonical space to effectively model temporal appearance variations. To prevent a decrease in the model's generalization ability to novel poses, we further propose low-dimensional global context to reduce unnecessary inter-body part dependencies and a quantization operation to mitigate overfitting of the delta pose sequence by the model. To validate the effectiveness of our approach, we collected a novel dataset named I3D-Human, with a focus on capturing temporal changes in clothing appearance under approximate poses. Through extensive experiments on both I3D-Human and existing datasets, our approach demonstrates superior qualitative and quantitative performance. In addition, our inertia-aware 3D human method can unprecedentedly simulate appearance changes caused by inertia at different velocities.
Authors: Chenyang Liu, Keyan Chen, Haotian Zhang, Zipeng Qi, Zhengxia Zou, Zhenwei Shi
Abstract: Monitoring changes in the Earth's surface is crucial for understanding natural processes and human impacts, necessitating precise and comprehensive interpretation methodologies. Remote sensing satellite imagery offers a unique perspective for monitoring these changes, leading to the emergence of remote sensing image change interpretation (RSICI) as a significant research focus. Current RSICI technology encompasses change detection and change captioning, each with its limitations in providing comprehensive interpretation. To address this, we propose an interactive Change-Agent, which can follow user instructions to achieve comprehensive change interpretation and insightful analysis, such as change detection and change captioning, change object counting, change cause analysis, etc. The Change-Agent integrates a multi-level change interpretation (MCI) model as the eyes and a large language model (LLM) as the brain. The MCI model contains two branches of pixel-level change detection and semantic-level change captioning, in which the BI-temporal Iterative Interaction (BI3) layer is proposed to enhance the model's discriminative feature representation capabilities. To support the training of the MCI model, we build the LEVIR-MCI dataset with a large number of change masks and captions of changes. Experiments demonstrate the SOTA performance of the MCI model in achieving both change detection and change description simultaneously, and highlight the promising application value of our Change-Agent in facilitating comprehensive interpretation of surface changes, which opens up a new avenue for intelligent remote sensing applications. To facilitate future research, we will make our dataset and codebase of the MCI model and Change-Agent publicly available at https://github.com/Chen-Yang-Liu/Change-Agent
Authors: Qianliang Wu, Haobo Jiang, Lei Luo, Jun Li, Yaqing Ding, Jin Xie, Jian Yang
Abstract: Establishing reliable correspondences is essential for registration tasks such as 3D and 2D3D registration. Existing methods commonly leverage geometric or semantic point features to generate potential correspondences. However, these features may face challenges such as large deformation, scale inconsistency, and ambiguous matching problems (e.g., symmetry). Additionally, many previous methods, which rely on single-pass prediction, may struggle with local minima in complex scenarios. To mitigate these challenges, we introduce a diffusion matching model for robust correspondence construction. Our approach treats correspondence estimation as a denoising diffusion process within the doubly stochastic matrix space, which gradually denoises (refines) a doubly stochastic matching matrix to the ground-truth one for high-quality correspondence estimation. It involves a forward diffusion process that gradually introduces Gaussian noise into the ground truth matching matrix and a reverse denoising process that iteratively refines the noisy matching matrix. In particular, the feature extraction from the backbone occurs only once during the inference phase. Our lightweight denoising module utilizes the same feature at each reverse sampling step. Evaluation of our method on both 3D and 2D3D registration tasks confirms its effectiveness. The code is available at https://github.com/wuqianliang/Diff-Reg.
Authors: Chenhao Xu, Chang-Tsun Li, Chee Peng Lim, Douglas Creighton
Abstract: Due to its deficiency in prior knowledge (inductive bias), Vision Transformer (ViT) requires pre-training on large-scale datasets to perform well. Moreover, the growing layers and parameters in ViT models impede their applicability to devices with limited computing resources. To mitigate the aforementioned challenges, this paper introduces a novel horizontally scalable vision transformer (HSViT) scheme. Specifically, a novel image-level feature embedding is introduced to ViT, where the preserved inductive bias allows the model to eliminate the need for pre-training while outperforming on small datasets. Besides, a novel horizontally scalable architecture is designed, facilitating collaborative model training and inference across multiple computing devices. The experimental results depict that, without pre-training, HSViT achieves up to 10% higher top-1 accuracy than state-of-the-art schemes on small datasets, while providing existing CNN backbones up to 3.1% improvement in top-1 accuracy on ImageNet. The code is available at https://github.com/xuchenhao001/HSViT.
Authors: Xiaofan Li, Zhizhong Zhang, Xin Tan, Chengwei Chen, Yanyun Qu, Yuan Xie, Lizhuang Ma
Abstract: The vision-language model has brought great improvement to few-shot industrial anomaly detection, which usually needs to design of hundreds of prompts through prompt engineering. For automated scenarios, we first use conventional prompt learning with many-class paradigm as the baseline to automatically learn prompts but found that it can not work well in one-class anomaly detection. To address the above problem, this paper proposes a one-class prompt learning method for few-shot anomaly detection, termed PromptAD. First, we propose semantic concatenation which can transpose normal prompts into anomaly prompts by concatenating normal prompts with anomaly suffixes, thus constructing a large number of negative samples used to guide prompt learning in one-class setting. Furthermore, to mitigate the training challenge caused by the absence of anomaly images, we introduce the concept of explicit anomaly margin, which is used to explicitly control the margin between normal prompt features and anomaly prompt features through a hyper-parameter. For image-level/pixel-level anomaly detection, PromptAD achieves first place in 11/12 few-shot settings on MVTec and VisA.
Authors: Maxence Bideaux, Alice Phe, Mohamed Chaouch, Bertrand Luvison, Quoc-Cuong Pham
Abstract: We introduce 3D-COCO, an extension of the original MS-COCO dataset providing 3D models and 2D-3D alignment annotations. 3D-COCO was designed to achieve computer vision tasks such as 3D reconstruction or image detection configurable with textual, 2D image, and 3D CAD model queries. We complete the existing MS-COCO dataset with 28K 3D models collected on ShapeNet and Objaverse. By using an IoU-based method, we match each MS-COCO annotation with the best 3D models to provide a 2D-3D alignment. The open-source nature of 3D-COCO is a premiere that should pave the way for new research on 3D-related topics. The dataset and its source codes is available at https://kalisteo.cea.fr/index.php/coco3d-object-detection-and-reconstruction/
URLs: https://kalisteo.cea.fr/index.php/coco3d-object-detection-and-reconstruction/
Authors: Heyuan Li, Ce Chen, Tianhao Shi, Yuda Qiu, Sizhe An, Guanying Chen, Xiaoguang Han
Abstract: While recent advances in 3D-aware Generative Adversarial Networks (GANs) have aided the development of near-frontal view human face synthesis, the challenge of comprehensively synthesizing a full 3D head viewable from all angles still persists. Although PanoHead proves the possibilities of using a large-scale dataset with images of both frontal and back views for full-head synthesis, it often causes artifacts for back views. Based on our in-depth analysis, we found the reasons are mainly twofold. First, from network architecture perspective, we found each plane in the utilized tri-plane/tri-grid representation space tends to confuse the features from both sides, causing "mirroring" artifacts (e.g., the glasses appear in the back). Second, from data supervision aspect, we found that existing discriminator training in 3D GANs mainly focuses on the quality of the rendered image itself, and does not care much about its plausibility with the perspective from which it was rendered. This makes it possible to generate "face" in non-frontal views, due to its easiness to fool the discriminator. In response, we propose SphereHead, a novel tri-plane representation in the spherical coordinate system that fits the human head's geometric characteristics and efficiently mitigates many of the generated artifacts. We further introduce a view-image consistency loss for the discriminator to emphasize the correspondence of the camera parameters and the images. The combination of these efforts results in visually superior outcomes with significantly fewer artifacts. Our code and dataset are publicly available at https://lhyfst.github.io/spherehead.
Authors: Yingshu Chen, Huajian Huang, Tuan-Anh Vu, Ka Chun Shum, Sai-Kit Yeung
Abstract: Creating large-scale virtual urban scenes with variant styles is inherently challenging. To facilitate prototypes of virtual production and bypass the need for complex materials and lighting setups, we introduce the first vision-and-text-driven texture stylization system for large-scale urban scenes, StyleCity. Taking an image and text as references, StyleCity stylizes a 3D textured mesh of a large-scale urban scene in a semantics-aware fashion and generates a harmonic omnidirectional sky background. To achieve that, we propose to stylize a neural texture field by transferring 2D vision-and-text priors to 3D globally and locally. During 3D stylization, we progressively scale the planned training views of the input 3D scene at different levels in order to preserve high-quality scene content. We then optimize the scene style globally by adapting the scale of the style image with the scale of the training views. Moreover, we enhance local semantics consistency by the semantics-aware style loss which is crucial for photo-realistic stylization. Besides texture stylization, we further adopt a generative diffusion model to synthesize a style-consistent omnidirectional sky image, which offers a more immersive atmosphere and assists the semantic stylization process. The stylized neural texture field can be baked into an arbitrary-resolution texture, enabling seamless integration into conventional rendering pipelines and significantly easing the virtual production prototyping process. Extensive experiments demonstrate our stylized scenes' superiority in qualitative and quantitative performance and user preferences.
Authors: Wonjae Kim, Sanghyuk Chun, Taekyung Kim, Dongyoon Han, Sangdoo Yun
Abstract: In an era where the volume of data drives the effectiveness of self-supervised learning, the specificity and clarity of data semantics play a crucial role in model training. Addressing this, we introduce HYPerbolic Entailment filtering (HYPE), a novel methodology designed to meticulously extract modality-wise meaningful and well-aligned data from extensive, noisy image-text pair datasets. Our approach leverages hyperbolic embeddings and the concept of entailment cones to evaluate and filter out samples with meaningless or underspecified semantics, focusing on enhancing the specificity of each data sample. HYPE not only demonstrates a significant improvement in filtering efficiency but also sets a new state-of-the-art in the DataComp benchmark when combined with existing filtering techniques. This breakthrough showcases the potential of HYPE to refine the data selection process, thereby contributing to the development of more accurate and efficient self-supervised learning models. Additionally, the image specificity $\epsilon_{i}$ can be independently applied to induce an image-only dataset from an image-text or image-only data pool for training image-only self-supervised models and showed superior performance when compared to the dataset induced by CLIP score.
Authors: Xuanchen Wang, Heng Wang, Dongnan Liu, Weidong Cai
Abstract: Automated choreography advances by generating dance from music. Current methods create skeleton keypoint sequences, not full dance videos, and cannot make specific individuals dance, limiting their real-world use. These methods also need precise keypoint annotations, making data collection difficult and restricting the use of self-made video datasets. To overcome these challenges, we introduce a novel task: generating dance videos directly from images of individuals guided by music. This task enables the dance generation of specific individuals without requiring keypoint annotations, making it more versatile and applicable to various situations. Our solution, the Dance Any Beat Diffusion model (DabFusion), utilizes a reference image and a music piece to generate dance videos featuring various dance types and choreographies. The music is analyzed by our specially designed music encoder, which identifies essential features including dance style, movement, and rhythm. DabFusion excels in generating dance videos not only for individuals in the training dataset but also for any previously unseen person. This versatility stems from its approach of generating latent optical flow, which contains all necessary motion information to animate any person in the image. We evaluate DabFusion's performance using the AIST++ dataset, focusing on video quality, audio-video synchronization, and motion-music alignment. We propose a 2D Motion-Music Alignment Score (2D-MM Align), which builds on the Beat Alignment Score to more effectively evaluate motion-music alignment for this new task. Experiments show that our DabFusion establishes a solid baseline for this innovative task. Video results can be found on our project page: https://DabFusion.github.io.
Authors: Haonan Zhang, Pengpeng Zeng, Lianli Gao, Jingkuan Song, Yihang Duan, Xinyu Lyu, Hengtao Shen
Abstract: Adapting large-scale image-text pre-training models, e.g., CLIP, to the video domain represents the current state-of-the-art for text-video retrieval. The primary approaches involve transferring text-video pairs to a common embedding space and leveraging cross-modal interactions on specific entities for semantic alignment. Though effective, these paradigms entail prohibitive computational costs, leading to inefficient retrieval. To address this, we propose a simple yet effective method, Global-Local Semantic Consistent Learning (GLSCL), which capitalizes on latent shared semantics across modalities for text-video retrieval. Specifically, we introduce a parameter-free global interaction module to explore coarse-grained alignment. Then, we devise a shared local interaction module that employs several learnable queries to capture latent semantic concepts for learning fine-grained alignment. Furthermore, an Inter-Consistency Loss (ICL) is devised to accomplish the concept alignment between the visual query and corresponding textual query, and an Intra-Diversity Loss (IDL) is developed to repulse the distribution within visual (textual) queries to generate more discriminative concepts. Extensive experiments on five widely used benchmarks (i.e., MSR-VTT, MSVD, DiDeMo, LSMDC, and ActivityNet) substantiate the superior effectiveness and efficiency of the proposed method. Remarkably, our method achieves comparable performance with SOTA as well as being nearly 220 times faster in terms of computational cost. Code is available at: https://github.com/zchoi/GLSCL.
Authors: Juntae Kim, Sungwon Woo, Jongho Nang
Abstract: Image copy detection is a task of detecting edited copies from any image within a reference database. While previous approaches have shown remarkable progress, the large size of their networks and descriptors remains disadvantage, complicating their practical application. In this paper, we propose a novel method that achieves a competitive performance by using a lightweight network and compact descriptors. By utilizing relational self-supervised distillation to transfer knowledge from a large network to a small network, we enable the training of lightweight networks with a small descriptor size. We introduce relational self-supervised distillation for flexible representation in a smaller feature space and applies contrastive learning with a hard negative loss to prevent dimensional collapse. For the DISC2021 benchmark, ResNet-50/EfficientNet-B0 are used as a teacher and student respectively, the micro average precision improved by 5.0%/4.9%/5.9% for 64/128/256 descriptor sizes compared to the baseline method.
Authors: Haoxing Chen, Yan Hong, Zizheng Huang, Zhuoer Xu, Zhangxuan Gu, Yaohui Li, Jun Lan, Huijia Zhu, Jianfu Zhang, Weiqiang Wang, Huaxiong Li
Abstract: Recently, video generation techniques have advanced rapidly. Given the popularity of video content on social media platforms, these models intensify concerns about the spread of fake information. Therefore, there is a growing demand for detectors capable of distinguishing between fake AI-generated videos and mitigating the potential harm caused by fake information. However, the lack of large-scale datasets from the most advanced video generators poses a barrier to the development of such detectors. To address this gap, we introduce the first AI-generated video detection dataset, GenVideo. It features the following characteristics: (1) a large volume of videos, including over one million AI-generated and real videos collected; (2) a rich diversity of generated content and methodologies, covering a broad spectrum of video categories and generation techniques. We conducted extensive studies of the dataset and proposed two evaluation methods tailored for real-world-like scenarios to assess the detectors' performance: the cross-generator video classification task assesses the generalizability of trained detectors on generators; the degraded video classification task evaluates the robustness of detectors to handle videos that have degraded in quality during dissemination. Moreover, we introduced a plug-and-play module, named Detail Mamba (DeMamba), designed to enhance the detectors by identifying AI-generated videos through the analysis of inconsistencies in temporal and spatial dimensions. Our extensive experiments demonstrate DeMamba's superior generalizability and robustness on GenVideo compared to existing detectors. We believe that the GenVideo dataset and the DeMamba module will significantly advance the field of AI-generated video detection. Our code and dataset will be aviliable at \url{https://github.com/chenhaoxing/DeMamba}.
Authors: Masashi Hatano, Ryo Hachiuma, Ryo Fujii, Hideo Saito
Abstract: We address a novel cross-domain few-shot learning task (CD-FSL) with multimodal input and unlabeled target data for egocentric action recognition. This paper simultaneously tackles two critical challenges associated with egocentric action recognition in CD-FSL settings: (1) the extreme domain gap in egocentric videos (e.g., daily life vs. industrial domain) and (2) the computational cost for real-world applications. We propose MM-CDFSL, a domain-adaptive and computationally efficient approach designed to enhance adaptability to the target domain and improve inference cost. To address the first challenge, we propose the incorporation of multimodal distillation into the student RGB model using teacher models. Each teacher model is trained independently on source and target data for its respective modality. Leveraging only unlabeled target data during multimodal distillation enhances the student model's adaptability to the target domain. We further introduce ensemble masked inference, a technique that reduces the number of input tokens through masking. In this approach, ensemble prediction mitigates the performance degradation caused by masking, effectively addressing the second issue. Our approach outperformed the state-of-the-art CD-FSL approaches with a substantial margin on multiple egocentric datasets, improving by an average of 6.12/6.10 points for 1-shot/5-shot settings while achieving $2.2$ times faster inference speed. Project page: https://masashi-hatano.github.io/MM-CDFSL/
Authors: Yuan Shen, Duygu Ceylan, Paul Guerrero, Zexiang Xu, Niloy J. Mitra, Shenlong Wang, Anna Fr\"uhst\"uck
Abstract: We present a simple, modular, and generic method that upsamples coarse 3D models by adding geometric and appearance details. While generative 3D models now exist, they do not yet match the quality of their counterparts in image and video domains. We demonstrate that it is possible to directly repurpose existing (pretrained) video models for 3D super-resolution and thus sidestep the problem of the shortage of large repositories of high-quality 3D training models. We describe how to repurpose video upsampling models, which are not 3D consistent, and combine them with 3D consolidation to produce 3D-consistent results. As output, we produce high quality Gaussian Splat models, which are object centric and effective. Our method is category agnostic and can be easily incorporated into existing 3D workflows. We evaluate our proposed SuperGaussian on a variety of 3D inputs, which are diverse both in terms of complexity and representation (e.g., Gaussian Splats or NeRFs), and demonstrate that our simple method significantly improves the fidelity of the final 3D models. Check our project website for details: supergaussian.github.io
Authors: Tianyuan Zhang, Lu Wang, Hainan Li, Yisong Xiao, Siyuan Liang, Aishan Liu, Xianglong Liu, Dacheng Tao
Abstract: Lane detection (LD) is an essential component of autonomous driving systems, providing fundamental functionalities like adaptive cruise control and automated lane centering. Existing LD benchmarks primarily focus on evaluating common cases, neglecting the robustness of LD models against environmental illusions such as shadows and tire marks on the road. This research gap poses significant safety challenges since these illusions exist naturally in real-world traffic situations. For the first time, this paper studies the potential threats caused by these environmental illusions to LD and establishes the first comprehensive benchmark LanEvil for evaluating the robustness of LD against this natural corruption. We systematically design 14 prevalent yet critical types of environmental illusions (e.g., shadow, reflection) that cover a wide spectrum of real-world influencing factors in LD tasks. Based on real-world environments, we create 94 realistic and customizable 3D cases using the widely used CARLA simulator, resulting in a dataset comprising 90,292 sampled images. Through extensive experiments, we benchmark the robustness of popular LD methods using LanEvil, revealing substantial performance degradation (-5.37% Accuracy and -10.70% F1-Score on average), with shadow effects posing the greatest risk (-7.39% Accuracy). Additionally, we assess the performance of commercial auto-driving systems OpenPilot and Apollo through collaborative simulations, demonstrating that proposed environmental illusions can lead to incorrect decisions and potential traffic accidents. To defend against environmental illusions, we propose the Attention Area Mixing (AAM) approach using hard examples, which witness significant robustness improvement (+3.76%) under illumination effects. We hope our paper can contribute to advancing more robust auto-driving systems in the future. Website: https://lanevil.github.io/.
Authors: Yuan-Ming Li, Wei-Jin Huang, An-Lan Wang, Ling-An Zeng, Jing-Ke Meng, Wei-Shi Zheng
Abstract: We present EgoExo-Fitness, a new full-body action understanding dataset, featuring fitness sequence videos recorded from synchronized egocentric and fixed exocentric (third-person) cameras. Compared with existing full-body action understanding datasets, EgoExo-Fitness not only contains videos from first-person perspectives, but also provides rich annotations. Specifically, two-level temporal boundaries are provided to localize single action videos along with sub-steps of each action. More importantly, EgoExo-Fitness introduces innovative annotations for interpretable action judgement--including technical keypoint verification, natural language comments on action execution, and action quality scores. Combining all of these, EgoExo-Fitness provides new resources to study egocentric and exocentric full-body action understanding across dimensions of "what", "when", and "how well". To facilitate research on egocentric and exocentric full-body action understanding, we construct benchmarks on a suite of tasks (i.e., action classification, action localization, cross-view sequence verification, cross-view skill determination, and a newly proposed task of guidance-based execution verification), together with detailed analysis. Code and data will be available at https://github.com/iSEE-Laboratory/EgoExo-Fitness/tree/main.
URLs: https://github.com/iSEE-Laboratory/EgoExo-Fitness/tree/main.
Authors: Pengfei Gu, Yejia Zhang, Huimin Li, Chaoli Wang, Danny Z. Chen
Abstract: Masked Autoencoders (MAEs) have been shown to be effective in pre-training Vision Transformers (ViTs) for natural and medical image analysis problems. By reconstructing missing pixel/voxel information in visible patches, a ViT encoder can aggregate contextual information for downstream tasks. But, existing MAE pre-training methods, which were specifically developed with the ViT architecture, lack the ability to capture geometric shape and spatial information, which is critical for medical image segmentation tasks. In this paper, we propose a novel extension of known MAEs for self pre-training (i.e., models pre-trained on the same target dataset) for 3D medical image segmentation. (1) We propose a new topological loss to preserve geometric shape information by computing topological signatures of both the input and reconstructed volumes, learning geometric shape information. (2) We introduce a pre-text task that predicts the positions of the centers and eight corners of 3D crops, enabling the MAE to aggregate spatial information. (3) We extend the MAE pre-training strategy to a hybrid state-of-the-art (SOTA) medical image segmentation architecture and co-pretrain it alongside the ViT. (4) We develop a fine-tuned model for downstream segmentation tasks by complementing the pre-trained ViT encoder with our pre-trained SOTA model. Extensive experiments on five public 3D segmentation datasets show the effectiveness of our new approach.
Authors: Jialong Wu, Mirko Meuter, Markus Schoeler, Matthias Rottmann
Abstract: Radar-based perception has gained increasing attention in autonomous driving, yet the inherent sparsity of radars poses challenges. Radar raw data often contains excessive noise, whereas radar point clouds retain only limited information. In this work, we holistically treat the sparse nature of radar data by introducing an adaptive subsampling method together with a tailored network architecture that exploits the sparsity patterns to discover global and local dependencies in the radar signal. Our subsampling module selects a subset of pixels from range-doppler (RD) spectra that contribute most to the downstream perception tasks. To improve the feature extraction on sparse subsampled data, we propose a new way of applying graph neural networks on radar data and design a novel two-branch backbone to capture both global and local neighbor information. An attentive fusion module is applied to combine features from both branches. Experiments on the RADIal dataset show that our SparseRadNet exceeds state-of-the-art (SOTA) performance in object detection and achieves close to SOTA accuracy in freespace segmentation, meanwhile using sparse subsampled input data.
Authors: Hejie Cui, Lingjun Mao, Xin Liang, Jieyu Zhang, Hui Ren, Quanzheng Li, Xiang Li, Carl Yang
Abstract: Recent advancements in multimodal foundation models have showcased impressive capabilities in understanding and reasoning with visual and textual information. Adapting these foundation models trained for general usage to specialized domains like biomedicine requires large-scale domain-specific instruction datasets. While existing works have explored curating such datasets automatically, the resultant datasets are not explicitly aligned with domain expertise. In this work, we propose a data-centric framework, Biomedical Visual Instruction Tuning with Clinician Preference Alignment (BioMed-VITAL), that incorporates clinician preferences into both stages of generating and selecting instruction data for tuning biomedical multimodal foundation models. First, during the generation stage, we prompt the GPT-4V generator with a diverse set of clinician-selected demonstrations for preference-aligned data candidate generation. Then, during the selection phase, we train a separate selection model, which explicitly distills clinician and policy-guided model preferences into a rating function to select high-quality data for medical instruction tuning. Results show that the model tuned with the instruction-following data from our method demonstrates a significant improvement in open visual chat (18.5% relatively) and medical VQA (win rate up to 81.73%). Our instruction-following data and models are available at BioMed-VITAL.github.io.
Authors: Wenxiao Cai, Yaroslav Ponomarenko, Jianhao Yuan, Xiaoqi Li, Wankou Yang, Hao Dong, Bo Zhao
Abstract: Vision Language Models (VLMs) have achieved impressive performance in 2D image understanding, however they are still struggling with spatial understanding which is the foundation of Embodied AI. In this paper, we propose SpatialBot for better spatial understanding by feeding both RGB and depth images. Additionally, we have constructed the SpatialQA dataset, which involves multi-level depth-related questions to train VLMs for depth understanding. Finally, we present SpatialBench to comprehensively evaluate VLMs' capabilities in spatial understanding at different levels. Extensive experiments on our spatial-understanding benchmark, general VLM benchmarks and Embodied AI tasks, demonstrate the remarkable improvements of SpatialBot trained on SpatialQA. The model, code and data are available at https://github.com/BAAI-DCAI/SpatialBot.
Authors: Bohan Li, Jiajun Deng, Wenyao Zhang, Zhujin Liang, Dalong Du, Xin Jin, Wenjun Zeng
Abstract: Camera-based 3D semantic scene completion (SSC) is pivotal for predicting complicated 3D layouts with limited 2D image observations. The existing mainstream solutions generally leverage temporal information by roughly stacking history frames to supplement the current frame, such straightforward temporal modeling inevitably diminishes valid clues and increases learning difficulty. To address this problem, we present HTCL, a novel Hierarchical Temporal Context Learning paradigm for improving camera-based semantic scene completion. The primary innovation of this work involves decomposing temporal context learning into two hierarchical steps: (a) cross-frame affinity measurement and (b) affinity-based dynamic refinement. Firstly, to separate critical relevant context from redundant information, we introduce the pattern affinity with scale-aware isolation and multiple independent learners for fine-grained contextual correspondence modeling. Subsequently, to dynamically compensate for incomplete observations, we adaptively refine the feature sampling locations based on initially identified locations with high affinity and their neighboring relevant regions. Our method ranks $1^{st}$ on the SemanticKITTI benchmark and even surpasses LiDAR-based methods in terms of mIoU on the OpenOccupancy benchmark. Our code is available on https://github.com/Arlo0o/HTCL.
Authors: Junsung Park, Kyungmin Kim, Hyunjung Shim
Abstract: Existing LiDAR semantic segmentation methods often struggle with performance declines in adverse weather conditions. Previous work has addressed this issue by simulating adverse weather or employing universal data augmentation during training. However, these methods lack a detailed analysis and understanding of how adverse weather negatively affects LiDAR semantic segmentation performance. Motivated by this issue, we identified key factors of adverse weather and conducted a toy experiment to pinpoint the main causes of performance degradation: (1) Geometric perturbation due to refraction caused by fog or droplets in the air and (2) Point drop due to energy absorption and occlusions. Based on these findings, we propose new strategic data augmentation techniques. First, we introduced a Selective Jittering (SJ) that jitters points in the random range of depth (or angle) to mimic geometric perturbation. Additionally, we developed a Learnable Point Drop (LPD) to learn vulnerable erase patterns with a Deep Q-Learning Network to approximate the point drop phenomenon from adverse weather conditions. Without precise weather simulation, these techniques strengthen the LiDAR semantic segmentation model by exposing it to vulnerable conditions identified by our data-centric analysis. Experimental results confirmed the suitability of the proposed data augmentation methods for enhancing robustness against adverse weather conditions. Our method achieves a notable 39.5 mIoU on the SemanticKITTI-to-SemanticSTF benchmark, improving the baseline by 8.1\%p and establishing a new state-of-the-art. Our code will be released at \url{https://github.com/engineerJPark/LiDARWeather}.
Authors: Prantik Howlader, Srijan Das, Hieu Le, Dimitris Samaras
Abstract: Incorporating pixel contextual information is critical for accurate segmentation. In this paper, we show that an effective way to incorporate contextual information is through a patch-based classifier. This patch classifier is trained to identify classes present within an image region, which facilitates the elimination of distractors and enhances the classification of small object segments. Specifically, we introduce Multi-scale Patch-based Multi-label Classifier (MPMC), a novel plug-in module designed for existing semi-supervised segmentation (SSS) frameworks. MPMC offers patch-level supervision, enabling the discrimination of pixel regions of different classes within a patch. Furthermore, MPMC learns an adaptive pseudo-label weight, using patch-level classification to alleviate the impact of the teacher's noisy pseudo-label supervision the student. This lightweight module can be integrated into any SSS framework, significantly enhancing their performance. We demonstrate the efficacy of our proposed MPMC by integrating it into four SSS methodologies and improving them across two natural image and one medical segmentation dataset, notably improving the segmentation results of the baselines across all the three datasets.
Authors: Anpei Chen, Haofei Xu, Stefano Esposito, Siyu Tang, Andreas Geiger
Abstract: Radiance field methods have achieved photorealistic novel view synthesis and geometry reconstruction. But they are mostly applied in per-scene optimization or small-baseline settings. While several recent works investigate feed-forward reconstruction with large baselines by utilizing transformers, they all operate with a standard global attention mechanism and hence ignore the local nature of 3D reconstruction. We propose a method that unifies local and global reasoning in transformer layers, resulting in improved quality and faster convergence. Our model represents scenes as Gaussian Volumes and combines this with an image encoder and Group Attention Layers for efficient feed-forward reconstruction. Experimental results demonstrate that our model, trained for two days on four GPUs, demonstrates high fidelity in reconstructing 360 deg radiance fields, and robustness to zero-shot and out-of-domain testing. Our project Page: https://apchenstu.github.io/LaRa/.
Authors: Brian Cheong, Jiachen Zhou, Steven Waslander
Abstract: Tracking-by-detection (TBD) methods achieve state-of-the-art performance on 3D tracking benchmarks for autonomous driving. On the other hand, tracking-by-attention (TBA) methods have the potential to outperform TBD methods, particularly for long occlusions and challenging detection settings. This work investigates why TBA methods continue to lag in performance behind TBD methods using a LiDAR-based joint detector and tracker called JDT3D. Based on this analysis, we propose two generalizable methods to bridge the gap between TBD and TBA methods: track sampling augmentation and confidence-based query propagation. JDT3D is trained and evaluated on the nuScenes dataset, achieving 0.574 on the AMOTA metric on the nuScenes test set, outperforming all existing LiDAR-based TBA approaches by over 6%. Based on our results, we further discuss some potential challenges with the existing TBA model formulation to explain the continued gap in performance with TBD methods. The implementation of JDT3D can be found at the following link: https://github.com/TRAILab/JDT3D.
Authors: Yuanhong Chen, Chong Wang, Yuyuan Liu, Hu Wang, Gustavo Carneiro
Abstract: Audio-visual segmentation (AVS) is an emerging task that aims to accurately segment sounding objects based on audio-visual cues. The success of AVS learning systems depends on the effectiveness of cross-modal interaction. Such a requirement can be naturally fulfilled by leveraging transformer-based segmentation architecture due to its inherent ability to capture long-range dependencies and flexibility in handling different modalities. However, the inherent training issues of transformer-based methods, such as the low efficacy of cross-attention and unstable bipartite matching, can be amplified in AVS, particularly when the learned audio query does not provide a clear semantic clue. In this paper, we address these two issues with the new Class-conditional Prompting Machine (CPM). CPM improves the bipartite matching with a learning strategy combining class-agnostic queries with class-conditional queries. The efficacy of cross-modal attention is upgraded with new learning objectives for the audio, visual and joint modalities. We conduct experiments on AVS benchmarks, demonstrating that our method achieves state-of-the-art (SOTA) segmentation accuracy.
Authors: Erdong Hu, Longteng Guo, Tongtian Yue, Zijia Zhao, Shuning Xue, Jing Liu
Abstract: In computer vision, Image Difference Captioning (IDC) is crucial for accurately describing variations between closely related images. Traditional IDC methods often rely on specialist models, which restrict their applicability across varied contexts. This paper introduces the OneDiff model, a novel generalist approach that utilizes a robust vision-language model architecture, integrating a siamese image encoder with a Visual Delta Module. This innovative configuration allows for the precise detection and articulation of fine-grained differences between image pairs. OneDiff is trained through a dual-phase strategy, encompassing Coupled Sample Training and multi-task learning across a diverse array of data types, supported by our newly developed DiffCap Dataset. This dataset merges real-world and synthetic data, enhancing the training process and bolstering the model's robustness. Extensive testing on diverse IDC benchmarks, such as Spot-the-Diff, CLEVR-Change, and Birds-to-Words, shows that OneDiff consistently outperforms existing state-of-the-art models in accuracy and adaptability, achieving improvements of up to 85\% CIDEr points in average. By setting a new benchmark in IDC, OneDiff paves the way for more versatile and effective applications in detecting and describing visual differences. The code, models, and data will be made publicly available.
Authors: Reza Abbasi, Mohammad Hossein Rohban, Mahdieh Soleymani Baghshah
Abstract: CLIP models have recently shown to exhibit Out of Distribution (OoD) generalization capabilities. However, Compositional Out of Distribution (C-OoD) generalization, which is a crucial aspect of a model's ability to understand unseen compositions of known concepts, is relatively unexplored for the CLIP models. Our goal is to address this problem and identify the factors that contribute to the C-OoD in CLIPs. We noted that previous studies regarding compositional understanding of CLIPs frequently fail to ensure that test samples are genuinely novel relative to the CLIP training data. To this end, we carefully synthesized a large and diverse dataset in the single object setting, comprising attributes for objects that are highly unlikely to be encountered in the combined training datasets of various CLIP models. This dataset enables an authentic evaluation of C-OoD generalization. Our observations reveal varying levels of C-OoD generalization across different CLIP models. We propose that the disentanglement of CLIP representations serves as a critical indicator in this context. By utilizing our synthesized datasets and other existing datasets, we assess various disentanglement metrics of text and image representations. Our study reveals that the disentanglement of image and text representations, particularly with respect to their compositional elements, plays a crucial role in improving the generalization of CLIP models in out-of-distribution settings. This finding suggests promising opportunities for advancing out-of-distribution generalization in CLIPs.
Authors: Jinhua Zhang, Hualian Sheng, Sijia Cai, Bing Deng, Qiao Liang, Wen Li, Ying Fu, Jieping Ye, Shuhang Gu
Abstract: Controllable generation is considered a potentially vital approach to address the challenge of annotating 3D data, and the precision of such controllable generation becomes particularly imperative in the context of data production for autonomous driving. Existing methods focus on the integration of diverse generative information into controlling inputs, utilizing frameworks such as GLIGEN or ControlNet, to produce commendable outcomes in controllable generation. However, such approaches intrinsically restrict generation performance to the learning capacities of predefined network architectures. In this paper, we explore the integration of controlling information and introduce PerlDiff (Perspective-Layout Diffusion Models), a method for effective street view image generation that fully leverages perspective 3D geometric information. Our PerlDiff employs 3D geometric priors to guide the generation of street view images with precise object-level control within the network learning process, resulting in a more robust and controllable output. Moreover, it demonstrates superior controllability compared to alternative layout control methods. Empirical results justify that our PerlDiff markedly enhances the precision of generation on the NuScenes and KITTI datasets. Our codes and models are publicly available at https://github.com/LabShuHangGU/PerlDiff.
Authors: Rui Li, Dong Liu
Abstract: Recent studies on motion estimation have advocated an optimized motion representation that is globally consistent across the entire video, preferably for every pixel. This is challenging as a uniform representation may not account for the complex and diverse motion and appearance of natural videos. We address this problem and propose a new test-time optimization method, named DecoMotion, for estimating per-pixel and long-range motion. DecoMotion explicitly decomposes video content into static scenes and dynamic objects, either of which uses a quasi-3D canonical volume to represent. DecoMotion separately coordinates the transformations between local and canonical spaces, facilitating an affine transformation for the static scene that corresponds to camera motion. For the dynamic volume, DecoMotion leverages discriminative and temporally consistent features to rectify the non-rigid transformation. The two volumes are finally fused to fully represent motion and appearance. This divide-and-conquer strategy leads to more robust tracking through occlusions and deformations and meanwhile obtains decomposed appearances. We conduct evaluations on the TAP-Vid benchmark. The results demonstrate our method boosts the point-tracking accuracy by a large margin and performs on par with some state-of-the-art dedicated point-tracking solutions.
Authors: Jinghang Li, Bangyan Liao, Xiuyuan LU, Peidong Liu, Shaojie Shen, Yi Zhou
Abstract: Predicting a potential collision with leading vehicles is an essential functionality of any autonomous/assisted driving system. One bottleneck of existing vision-based solutions is that their updating rate is limited to the frame rate of standard cameras used. In this paper, we present a novel method that estimates the time to collision using a neuromorphic event-based camera, a biologically inspired visual sensor that can sense at exactly the same rate as scene dynamics. The core of the proposed algorithm consists of a two-step approach for efficient and accurate geometric model fitting on event data in a coarse-to-fine manner. The first step is a robust linear solver based on a novel geometric measurement that overcomes the partial observability of event-based normal flow. The second step further refines the resulting model via a spatio-temporal registration process formulated as a nonlinear optimization problem. Experiments on both synthetic and real data demonstrate the effectiveness of the proposed method, outperforming other alternative methods in terms of efficiency and accuracy.
Authors: Xianghao Kong, Jinyu Chen, Wenguan Wang, Hang Su, Xiaolin Hu, Yi Yang, Si Liu
Abstract: Instruction generation is a vital and multidisciplinary research area with broad applications. Existing instruction generation models are limited to generating instructions in a single style from a particular dataset, and the style and content of generated instructions cannot be controlled. Moreover, most existing instruction generation methods also disregard the spatial modeling of the navigation environment. Leveraging the capabilities of Large Language Models (LLMs), we propose C-Instructor, which utilizes the chain-of-thought-style prompt for style-controllable and content-controllable instruction generation. Firstly, we propose a Chain of Thought with Landmarks (CoTL) mechanism, which guides the LLM to identify key landmarks and then generate complete instructions. CoTL renders generated instructions more accessible to follow and offers greater controllability over the manipulation of landmark objects. Furthermore, we present a Spatial Topology Modeling Task to facilitate the understanding of the spatial structure of the environment. Finally, we introduce a Style-Mixed Training policy, harnessing the prior knowledge of LLMs to enable style control for instruction generation based on different prompts within a single model instance. Extensive experiments demonstrate that instructions generated by C-Instructor outperform those generated by previous methods in text metrics, navigation guidance evaluation, and user studies.
Authors: Minghui Wu, Chenxu Zhao, Anyang Su, Donglin Di, Tianyu Fu, Da An, Min He, Ya Gao, Meng Ma, Kun Yan, Ping Wang
Abstract: Understanding of video creativity and content often varies among individuals, with differences in focal points and cognitive levels across different ages, experiences, and genders. There is currently a lack of research in this area, and most existing benchmarks suffer from several drawbacks: 1) a limited number of modalities and answers with restrictive length; 2) the content and scenarios within the videos are excessively monotonous, transmitting allegories and emotions that are overly simplistic. To bridge the gap to real-world applications, we introduce a large-scale Subjective Response Indicators for Advertisement Videos dataset, namely SRI-ADV. Specifically, we collected real changes in Electroencephalographic (EEG) and eye-tracking regions from different demographics while they viewed identical video content. Utilizing this multi-modal dataset, we developed tasks and protocols to analyze and evaluate the extent of cognitive understanding of video content among different users. Along with the dataset, we designed a Hypergraph Multi-modal Large Language Model (HMLLM) to explore the associations among different demographics, video elements, EEG, and eye-tracking indicators. HMLLM could bridge semantic gaps across rich modalities and integrate information beyond different modalities to perform logical reasoning. Extensive experimental evaluations on SRI-ADV and other additional video-based generative performance benchmarks demonstrate the effectiveness of our method. The codes and dataset will be released at https://github.com/suay1113/HMLLM.
Authors: Zhengbo Zhang, Li Xu, Duo Peng, Hossein Rahmani, Jun Liu
Abstract: We introduce Diff-Tracker, a novel approach for the challenging unsupervised visual tracking task leveraging the pre-trained text-to-image diffusion model. Our main idea is to leverage the rich knowledge encapsulated within the pre-trained diffusion model, such as the understanding of image semantics and structural information, to address unsupervised visual tracking. To this end, we design an initial prompt learner to enable the diffusion model to recognize the tracking target by learning a prompt representing the target. Furthermore, to facilitate dynamic adaptation of the prompt to the target's movements, we propose an online prompt updater. Extensive experiments on five benchmark datasets demonstrate the effectiveness of our proposed method, which also achieves state-of-the-art performance.
Authors: Tahira Shehzadi, Ifza, Didier Stricker, Muhammad Zeshan Afzal
Abstract: The impressive advancements in semi-supervised learning have driven researchers to explore its potential in object detection tasks within the field of computer vision. Semi-Supervised Object Detection (SSOD) leverages a combination of a small labeled dataset and a larger, unlabeled dataset. This approach effectively reduces the dependence on large labeled datasets, which are often expensive and time-consuming to obtain. Initially, SSOD models encountered challenges in effectively leveraging unlabeled data and managing noise in generated pseudo-labels for unlabeled data. However, numerous recent advancements have addressed these issues, resulting in substantial improvements in SSOD performance. This paper presents a comprehensive review of 27 cutting-edge developments in SSOD methodologies, from Convolutional Neural Networks (CNNs) to Transformers. We delve into the core components of semi-supervised learning and its integration into object detection frameworks, covering data augmentation techniques, pseudo-labeling strategies, consistency regularization, and adversarial training methods. Furthermore, we conduct a comparative analysis of various SSOD models, evaluating their performance and architectural differences. We aim to ignite further research interest in overcoming existing challenges and exploring new directions in semi-supervised learning for object detection.
Authors: Morris Alper, Hadar Averbuch-Elor
Abstract: While recent vision-and-language models (VLMs) like CLIP are a powerful tool for analyzing text and images in a shared semantic space, they do not explicitly model the hierarchical nature of the set of texts which may describe an image. Conversely, existing multimodal hierarchical representation learning methods require costly training from scratch, failing to leverage the knowledge encoded by state-of-the-art multimodal foundation models. In this work, we study the knowledge of existing foundation models, finding that they exhibit emergent understanding of visual-semantic hierarchies despite not being directly trained for this purpose. We propose the Radial Embedding (RE) framework for probing and optimizing hierarchical understanding, and contribute the HierarCaps dataset, a benchmark facilitating the study of hierarchical knowledge in image--text representations, constructed automatically via large language models. Our results show that foundation VLMs exhibit zero-shot hierarchical understanding, surpassing the performance of prior models explicitly designed for this purpose. Furthermore, we show that foundation models may be better aligned to hierarchical reasoning via a text-only fine-tuning phase, while retaining pretraining knowledge.
Authors: Weixiang Sun, Xiaocao You, Ruizhe Zheng, Zhengqing Yuan, Xiang Li, Lifang He, Quanzheng Li, Lichao Sun
Abstract: Generative models hold promise for revolutionizing medical education, robot-assisted surgery, and data augmentation for medical AI development. Diffusion models can now generate realistic images from text prompts, while recent advancements have demonstrated their ability to create diverse, high-quality videos. However, these models often struggle with generating accurate representations of medical procedures and detailed anatomical structures. This paper introduces Bora, the first spatio-temporal diffusion probabilistic model designed for text-guided biomedical video generation. Bora leverages Transformer architecture and is pre-trained on general-purpose video generation tasks. It is fine-tuned through model alignment and instruction tuning using a newly established medical video corpus, which includes paired text-video data from various biomedical fields. To the best of our knowledge, this is the first attempt to establish such a comprehensive annotated biomedical video dataset. Bora is capable of generating high-quality video data across four distinct biomedical domains, adhering to medical expert standards and demonstrating consistency and diversity. This generalist video generative model holds significant potential for enhancing medical consultation and decision-making, particularly in resource-limited settings. Additionally, Bora could pave the way for immersive medical training and procedure planning. Extensive experiments on distinct medical modalities such as endoscopy, ultrasound, MRI, and cell tracking validate the effectiveness of our model in understanding biomedical instructions and its superior performance across subjects compared to state-of-the-art generation models.
Authors: Milena T. Bagdasarian, Paul Knoll, Florian Barthel, Anna Hilsmann, Peter Eisert, Wieland Morgenstern
Abstract: We present a work-in-progress survey on 3D Gaussian Splatting compression methods, focusing on their statistical performance across various benchmarks. This survey aims to facilitate comparability by summarizing key statistics of different compression approaches in a tabulated format. The datasets evaluated include TanksAndTemples, MipNeRF360, DeepBlending, and SyntheticNeRF. For each method, we report the Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), Learned Perceptual Image Patch Similarity (LPIPS), and the resultant size in megabytes (MB), as provided by the respective authors. This is an ongoing, open project, and we invite contributions from the research community as GitHub issues or pull requests. Please visit http://w-m.github.io/3dgs-compression-survey/ for more information and a sortable version of the table.
Authors: Lorenzo Vaquero, Yihong Xu, Xavier Alameda-Pineda, Victor M. Brea, Manuel Mucientes
Abstract: Multi-object tracking (MOT) endeavors to precisely estimate the positions and identities of multiple objects over time. The prevailing approach, tracking-by-detection (TbD), first detects objects and then links detections, resulting in a simple yet effective method. However, contemporary detectors may occasionally miss some objects in certain frames, causing trackers to cease tracking prematurely. To tackle this issue, we propose BUSCA, meaning `to search', a versatile framework compatible with any online TbD system, enhancing its ability to persistently track those objects missed by the detector, primarily due to occlusions. Remarkably, this is accomplished without modifying past tracking results or accessing future frames, i.e., in a fully online manner. BUSCA generates proposals based on neighboring tracks, motion, and learned tokens. Utilizing a decision Transformer that integrates multimodal visual and spatiotemporal information, it addresses the object-proposal association as a multi-choice question-answering task. BUSCA is trained independently of the underlying tracker, solely on synthetic data, without requiring fine-tuning. Through BUSCA, we showcase consistent performance enhancements across five different trackers and establish a new state-of-the-art baseline across three different benchmarks. Code available at: https://github.com/lorenzovaquero/BUSCA.
Authors: Haonan Wang, Jie Liu, Jie Tang, Gangshan Wu, Bo Xu, Yanbing Chou, Yong Wang
Abstract: In recent years, 2D human pose estimation has made significant progress on public benchmarks. However, many of these approaches face challenges of less applicability in the industrial community due to the large number of parametric quantities and computational overhead. Efficient human pose estimation remains a hurdle, especially for whole-body pose estimation with numerous keypoints. While most current methods for efficient human pose estimation primarily rely on CNNs, we propose the Group-based Token Pruning Transformer (GTPT) that fully harnesses the advantages of the Transformer. GTPT alleviates the computational burden by gradually introducing keypoints in a coarse-to-fine manner. It minimizes the computation overhead while ensuring high performance. Besides, GTPT groups keypoint tokens and prunes visual tokens to improve model performance while reducing redundancy. We propose the Multi-Head Group Attention (MHGA) between different groups to achieve global interaction with little computational overhead. We conducted experiments on COCO and COCO-WholeBody. Compared to other methods, the experimental results show that GTPT can achieve higher performance with less computation, especially in whole-body with numerous keypoints.
Authors: Jae Myung Kim, Jessica Bader, Stephan Alaniz, Cordelia Schmid, Zeynep Akata
Abstract: While text-to-image diffusion models have been shown to achieve state-of-the-art results in image synthesis, they have yet to prove their effectiveness in downstream applications. Previous work has proposed to generate data for image classifier training given limited real data access. However, these methods struggle to generate in-distribution images or depict fine-grained features, thereby hindering the generalization of classification models trained on synthetic datasets. We propose DataDream, a framework for synthesizing classification datasets that more faithfully represents the real data distribution when guided by few-shot examples of the target classes. DataDream fine-tunes LoRA weights for the image generation model on the few real images before generating the training data using the adapted model. We then fine-tune LoRA weights for CLIP using the synthetic data to improve downstream image classification over previous approaches on a large variety of datasets. We demonstrate the efficacy of DataDream through extensive experiments, surpassing state-of-the-art classification accuracy with few-shot data across 7 out of 10 datasets, while being competitive on the other 3. Additionally, we provide insights into the impact of various factors, such as the number of real-shot and generated images as well as the fine-tuning compute on model performance. The code is available at https://github.com/ExplainableML/DataDream.
Authors: Aishwarya Jadhav
Abstract: Treatment of stagnant water bodies that act as a breeding site for malarial vectors is a fundamental step in most malaria elimination campaigns. However, identification of such water bodies over large areas is expensive, labour-intensive and time-consuming and hence, challenging in countries with limited resources. Practical models that can efficiently locate water bodies can target the limited resources by greatly reducing the area that needs to be scanned by the field workers. To this end, we propose a practical topographic model based on easily available, global, high-resolution DEM data to predict locations of potential vector-breeding water sites. We surveyed the Obuasi region of Ghana to assess the impact of various topographic features on different types of water bodies and uncover the features that significantly influence the formation of aquatic habitats. We further evaluate the effectiveness of multiple models. Our best model significantly outperforms earlier attempts that employ topographic variables for detection of small water sites, even the ones that utilize additional satellite imagery data and demonstrates robustness across different settings.
Authors: Lukas Struppek, Dominik Hintersdorf, Daniel Neider, Kristian Kersting
Abstract: Apple recently revealed its deep perceptual hashing system NeuralHash to detect child sexual abuse material (CSAM) on user devices before files are uploaded to its iCloud service. Public criticism quickly arose regarding the protection of user privacy and the system's reliability. In this paper, we present the first comprehensive empirical analysis of deep perceptual hashing based on NeuralHash. Specifically, we show that current deep perceptual hashing may not be robust. An adversary can manipulate the hash values by applying slight changes in images, either induced by gradient-based approaches or simply by performing standard image transformations, forcing or preventing hash collisions. Such attacks permit malicious actors easily to exploit the detection system: from hiding abusive material to framing innocent users, everything is possible. Moreover, using the hash values, inferences can still be made about the data stored on user devices. In our view, based on our results, deep perceptual hashing in its current form is generally not ready for robust client-side scanning and should not be used from a privacy perspective.
Authors: Neil Mallinar, James B. Simon, Amirhesam Abedsoltan, Parthe Pandit, Mikhail Belkin, Preetum Nakkiran
Abstract: The practical success of overparameterized neural networks has motivated the recent scientific study of interpolating methods, which perfectly fit their training data. Certain interpolating methods, including neural networks, can fit noisy training data without catastrophically bad test performance, in defiance of standard intuitions from statistical learning theory. Aiming to explain this, a body of recent work has studied benign overfitting, a phenomenon where some interpolating methods approach Bayes optimality, even in the presence of noise. In this work we argue that while benign overfitting has been instructive and fruitful to study, many real interpolating methods like neural networks do not fit benignly: modest noise in the training set causes nonzero (but non-infinite) excess risk at test time, implying these models are neither benign nor catastrophic but rather fall in an intermediate regime. We call this intermediate regime tempered overfitting, and we initiate its systematic study. We first explore this phenomenon in the context of kernel (ridge) regression (KR) by obtaining conditions on the ridge parameter and kernel eigenspectrum under which KR exhibits each of the three behaviors. We find that kernels with powerlaw spectra, including Laplace kernels and ReLU neural tangent kernels, exhibit tempered overfitting. We then empirically study deep neural networks through the lens of our taxonomy, and find that those trained to interpolation are tempered, while those stopped early are benign. We hope our work leads to a more refined understanding of overfitting in modern learning.
Authors: Olivier Rukundo
Abstract: Current mask processing operations rely on interpolation algorithms that do not produce extra pixels, such as nearest neighbor (NN) interpolation, as opposed to algorithms that do produce extra pixels, like bicubic (BIC) or bilinear (BIL) interpolation. In our previous study, the author proposed an alternative approach to NN-based mask processing and evaluated its effects on deep learning training outcomes. In this study, the author evaluated the effects of both BIC-based image and mask processing and BIC-and-NN-based image and mask processing versus NN-based image and mask processing. The evaluation revealed that the BIC-BIC model/network was an 8.9578 % (with image size 256 x 256) and a 1.0496 % (with image size 384 x 384) increase of the NN-NN network compared to the NN-BIC network which was an 8.3127 % (with image size 256 x 256) and a 0.2887 % (with image size 384 x 384) increase of the NN-NN network.
Authors: Hongtao Kang, Die Luo, Li Chen, Junbo Hu, Tingwei Quan, Shaoqun Zeng, Shenghua Cheng, Xiuli Liu
Abstract: In practice, digital pathology images are often affected by various factors, resulting in very large differences in color and brightness. Stain normalization can effectively reduce the differences in color and brightness of digital pathology images, thus improving the performance of computer-aided diagnostic systems. Conventional stain normalization methods rely on one or several reference images, but one or several images may not adequately represent the entire dataset. Although learning-based stain normalization methods are a general approach, they use complex deep networks, which not only greatly reduce computational efficiency, but also risk introducing artifacts. Some studies use specialized network structures to enhance computational efficiency and reliability, but these methods are difficult to apply to multi-to-one stain normalization due to insufficient network capacity. In this study, we introduced dynamic-parameter network and proposed a novel method for stain normalization, called ParamNet. ParamNet addresses the challenges of limited network capacity and computational efficiency by introducing dynamic parameters (weights and biases of convolutional layers) into the network design. By effectively leveraging these parameters, ParamNet achieves superior performance in stain normalization while maintaining computational efficiency. Results show ParamNet can normalize one whole slide image (WSI) of 100,000x100,000 within 25s. The code is available at: https://github.com/khtao/ParamNet.
Authors: Kirill Aistov, Maxim Koroteev
Abstract: Based on the standard VMAF implementation we propose an implementation of VMAF using PyTorch framework. For this implementation comparisons with the standard (libvmaf) show the discrepancy $\lesssim 10^{-2}$ in VMAF units. We investigate gradients computation when using VMAF as an objective function and demonstrate that training using this function does not result in ill-behaving gradients. The implementation is then used to train a preprocessing filter. It is demonstrated that its performance is superior to the unsharp masking filter. The resulting filter is also easy for implementation and can be applied in video processing tasks for video copression improvement. This is confirmed by the results of numerical experiments.
Authors: Guanchu Wang, Yu-Neng Chuang, Fan Yang, Mengnan Du, Chia-Yuan Chang, Shaochen Zhong, Zirui Liu, Zhaozhuo Xu, Kaixiong Zhou, Xuanting Cai, Xia Hu
Abstract: Explainable machine learning significantly improves the transparency of deep neural networks. However, existing work is constrained to explaining the behavior of individual model predictions, and lacks the ability to transfer the explanation across various models and tasks. This limitation results in explaining various tasks being time- and resource-consuming. To address this problem, we introduce a Transferable Vision Explainer (TVE) that can effectively explain various vision models in downstream tasks. Specifically, the transferability of TVE is realized through a pre-training process on large-scale datasets towards learning the meta-attribution. This meta-attribution leverages the versatility of generic backbone encoders to comprehensively encode the attribution knowledge for the input instance, which enables TVE to seamlessly transfer to explain various downstream tasks, without the need for training on task-specific data. Empirical studies involve explaining three different architectures of vision models across three diverse downstream datasets. The experimental results indicate TVE is effective in explaining these tasks without the need for additional training on downstream data.
Authors: Md Rifat Arefin, Yan Zhang, Aristide Baratin, Francesco Locatello, Irina Rish, Dianbo Liu, Kenji Kawaguchi
Abstract: Models prone to spurious correlations in training data often produce brittle predictions and introduce unintended biases. Addressing this challenge typically involves methods relying on prior knowledge and group annotation to remove spurious correlations, which may not be readily available in many applications. In this paper, we establish a novel connection between unsupervised object-centric learning and mitigation of spurious correlations. Instead of directly inferring subgroups with varying correlations with labels, our approach focuses on discovering concepts: discrete ideas that are shared across input samples. Leveraging existing object-centric representation learning, we introduce CoBalT: a concept balancing technique that effectively mitigates spurious correlations without requiring human labeling of subgroups. Evaluation across the benchmark datasets for sub-population shifts demonstrate superior or competitive performance compared state-of-the-art baselines, without the need for group annotation. Code is available at https://github.com/rarefin/CoBalT.
Authors: Zhennong Chen, Sekeun Kim, Hui Ren, Quanzheng Li, Xiang Li
Abstract: Accurate myocardium segmentation across all phases in one cardiac cycle in cine cardiac magnetic resonance (CMR) scans is crucial for comprehensively cardiac function analysis. Despite advancements in deep learning (DL) for automatic cine CMR segmentation, generalizability on unseen data remains a significant challenge. Recently, the segment-anything-model (SAM) has been invented as a segmentation foundation model, known for its accurate segmentation and more importantly, zero-shot generalization. SAM was trained on two-dimensional (2D) natural images; to adapt it for comprehensive cine CMR segmentation, we propose cineCMR-SAM which incorporates both temporal and spatial information through a modified model architecture. Compared to other state-of-the-art (SOTA) methods, our model achieved superior data-specific model segmentation accuracy on the STACOM2011 when fine-tuned on this dataset and demonstrated superior zero-shot generalization on two other large public datasets (ACDC and M&Ms) unseen during fine-tuning. Additionally, we introduced a text prompt feature in cineCMR-SAM to specify the view type of input slices (short-axis or long-axis), enhancing performance across all view types.
Authors: Xinru Zhang, Ni Ou, Berke Doga Basaran, Marco Visentin, Mengyun Qiao, Renyang Gu, Cheng Ouyang, Yaou Liu, Paul M. Matthew, Chuyang Ye, Wenjia Bai
Abstract: Brain lesion segmentation plays an essential role in neurological research and diagnosis. As brain lesions can be caused by various pathological alterations, different types of brain lesions tend to manifest with different characteristics on different imaging modalities. Due to this complexity, brain lesion segmentation methods are often developed in a task-specific manner. A specific segmentation model is developed for a particular lesion type and imaging modality. However, the use of task-specific models requires predetermination of the lesion type and imaging modality, which complicates their deployment in real-world scenarios. In this work, we propose a universal foundation model for 3D brain lesion segmentation, which can automatically segment different types of brain lesions for input data of various imaging modalities. We formulate a novel Mixture of Modality Experts (MoME) framework with multiple expert networks attending to different imaging modalities. A hierarchical gating network combines the expert predictions and fosters expertise collaboration. Furthermore, we introduce a curriculum learning strategy during training to avoid the degeneration of each expert network and preserve their specialization. We evaluated the proposed method on nine brain lesion datasets, encompassing five imaging modalities and eight lesion types. The results show that our model outperforms state-of-the-art universal models and provides promising generalization to unseen datasets.
Authors: Yoeri Poels, Koen Minartz, Harshit Bansal, Vlado Menkovski
Abstract: Simulation is a powerful tool to better understand physical systems, but generally requires computationally expensive numerical methods. Downstream applications of such simulations can become computationally infeasible if they require many forward solves, for example in the case of inverse design with many degrees of freedom. In this work, we investigate and extend neural PDE solvers as a tool to aid in scaling simulations for two-phase flow problems, and simulations of oil expulsion from a pore specifically. We extend existing numerical methods for this problem to a more complex setting involving varying geometries of the domain to generate a challenging dataset. Further, we investigate three prominent neural PDE solver methods, namely the UNet, DRN, and U-FNO, and extend them for characteristics of the oil-expulsion problem: (1) spatial conditioning on the geometry; (2) periodicity in the boundary; (3) approximate mass conservation. We scale all methods and benchmark their speed-accuracy trade-off, evaluate qualitative properties, and perform an ablation study. We find that the investigated methods can accurately model the droplet dynamics with up to three orders of magnitude speed-up, that our extensions improve performance over the baselines, and that the introduced varying geometries constitute a significantly more challenging setting over the previously considered oil expulsion problem.
Authors: Abdulaziz Almuzairee, Nicklas Hansen, Henrik I. Christensen
Abstract: Q-learning algorithms are appealing for real-world applications due to their data-efficiency, but they are very prone to overfitting and training instabilities when trained from visual observations. Prior work, namely SVEA, finds that selective application of data augmentation can improve the visual generalization of RL agents without destabilizing training. We revisit its recipe for data augmentation, and find an assumption that limits its effectiveness to augmentations of a photometric nature. Addressing these limitations, we propose a generalized recipe, SADA, that works with wider varieties of augmentations. We benchmark its effectiveness on DMC-GB2 - our proposed extension of the popular DMControl Generalization Benchmark - as well as tasks from Meta-World and the Distracting Control Suite, and find that our method, SADA, greatly improves training stability and generalization of RL agents across a diverse set of augmentations. For visualizations, code and benchmark: see https://aalmuzairee.github.io/SADA/
Authors: Detian Chu, Linyuan Bai, Jianuo Huang, Zhenlong Fang, Peng Zhang, Wei Kang
Abstract: With the advancement of autonomous driving, ensuring safety during motion planning and navigation is becoming more and more important. However, most end-to-end planning methods suffer from a lack of safety. This research addresses the safety issue in the control optimization problem of autonomous driving, formulated as Constrained Markov Decision Processes (CMDPs). We propose a novel, model-based approach for policy optimization, utilizing a conditional Value-at-Risk based Soft Actor Critic to manage constraints in complex, high-dimensional state spaces effectively. Our method introduces a worst-case actor to guide safe exploration, ensuring rigorous adherence to safety requirements even in unpredictable scenarios. The policy optimization employs the Augmented Lagrangian method and leverages latent diffusion models to predict and simulate future trajectories. This dual approach not only aids in navigating environments safely but also refines the policy's performance by integrating distribution modeling to account for environmental uncertainties. Empirical evaluations conducted in both simulated and real environment demonstrate that our approach outperforms existing methods in terms of safety, efficiency, and decision-making capabilities.
Authors: Quang H. Nguyen, Nguyen Ngoc-Hieu, The-Anh Ta, Thanh Nguyen-Tang, Kok-Seng Wong, Hoang Thanh-Tung, Khoa D. Doan
Abstract: Deep neural networks are vulnerable to backdoor attacks, a type of adversarial attack that poisons the training data to manipulate the behavior of models trained on such data. Clean-label attacks are a more stealthy form of backdoor attacks that can perform the attack without changing the labels of poisoned data. Early works on clean-label attacks added triggers to a random subset of the training set, ignoring the fact that samples contribute unequally to the attack's success. This results in high poisoning rates and low attack success rates. To alleviate the problem, several supervised learning-based sample selection strategies have been proposed. However, these methods assume access to the entire labeled training set and require training, which is expensive and may not always be practical. This work studies a new and more practical (but also more challenging) threat model where the attacker only provides data for the target class (e.g., in face recognition systems) and has no knowledge of the victim model or any other classes in the training set. We study different strategies for selectively poisoning a small set of training samples in the target class to boost the attack success rate in this setting. Our threat model poses a serious threat in training machine learning models with third-party datasets, since the attack can be performed effectively with limited information. Experiments on benchmark datasets illustrate the effectiveness of our strategies in improving clean-label backdoor attacks.