Authors: Samy Blusseau (CMM)
Abstract: Morphological neural networks, or layers, can be a powerful tool to boost the progress in mathematical morphology, either on theoretical aspects such as the representation of complete lattice operators, or in the development of image processing pipelines. However, these architectures turn out to be difficult to train when they count more than a few morphological layers, at least within popular machine learning frameworks which use gradient descent based optimization algorithms. In this paper we investigate the potential and limitations of differentiation based approaches and back-propagation applied to morphological networks, in light of the non-smooth optimization concept of Bouligand derivative. We provide insights and first theoretical guidelines, in particular regarding initialization and learning rates.
Authors: Lachlan Thorpe, Lewis Bawden, Karanjot Vendal, John Bronskill, Richard E. Turner
Abstract: We present a transformer decoder based model, SportsNGEN, that is trained on sports player and ball tracking sequences that is capable of generating realistic and sustained gameplay. We train and evaluate SportsNGEN on a large database of professional tennis tracking data and demonstrate that by combining the generated simulations with a shot classifier and logic to start and end rallies, the system is capable of simulating an entire tennis match. In addition, a generic version of SportsNGEN can be customized to a specific player by fine-tuning on match data that includes that player. We show that our model is well calibrated and can be used to derive insights for coaches and broadcasters by evaluating counterfactual or what if options. Finally, we show qualitative results indicating the same approach works for football.
Authors: Xiang Meng, Shibal Ibrahim, Kayhan Behdin, Hussein Hazimeh, Natalia Ponomareva, Rahul Mazumder
Abstract: Structured pruning is a promising approach for reducing the inference costs of large vision and language models. By removing carefully chosen structures, e.g., neurons or attention heads, the improvements from this approach can be realized on standard deep learning hardware. In this work, we focus on structured pruning in the one-shot (post-training) setting, which does not require model retraining after pruning. We propose a novel combinatorial optimization framework for this problem, based on a layer-wise reconstruction objective and a careful reformulation that allows for scalable optimization. Moreover, we design a new local combinatorial optimization algorithm, which exploits low-rank updates for efficient local search. Our framework is time and memory-efficient and considerably improves upon state-of-the-art one-shot methods on vision models (e.g., ResNet50, MobileNet) and language models (e.g., OPT-1.3B -- OPT-30B). For language models, e.g., OPT-2.7B, OSSCAR can lead to $125\times$ lower test perplexity on WikiText with $2\times$ inference time speedup in comparison to the state-of-the-art ZipLM approach. Our framework is also $6\times$ -- $8\times$ faster. Notably, our work considers models with tens of billions of parameters, which is up to $100\times$ larger than what has been previously considered in the structured pruning literature.
Authors: Qianhan Feng, Lujing Xie, Shijie Fang, Tong Lin
Abstract: Semi-supervised Learning (SSL) reduces the need for extensive annotations in deep learning, but the more realistic challenge of imbalanced data distribution in SSL remains largely unexplored. In Class Imbalanced Semi-supervised Learning (CISSL), the bias introduced by unreliable pseudo-labels can be exacerbated by imbalanced data distributions. Most existing methods address this issue at instance-level through reweighting or resampling, but the performance is heavily limited by their reliance on biased backbone representation. Some other methods do perform feature-level adjustments like feature blending but might introduce unfavorable noise. In this paper, we discuss the bonus of a more balanced feature distribution for the CISSL problem, and further propose a Balanced Feature-Level Contrastive Learning method (BaCon). Our method directly regularizes the distribution of instances' representations in a well-designed contrastive manner. Specifically, class-wise feature centers are computed as the positive anchors, while negative anchors are selected by a straightforward yet effective mechanism. A distribution-related temperature adjustment is leveraged to control the class-wise contrastive degrees dynamically. Our method demonstrates its effectiveness through comprehensive experiments on the CIFAR10-LT, CIFAR100-LT, STL10-LT, and SVHN-LT datasets across various settings. For example, BaCon surpasses instance-level method FixMatch-based ABC on CIFAR10-LT with a 1.21% accuracy improvement, and outperforms state-of-the-art feature-level method CoSSL on CIFAR100-LT with a 0.63% accuracy improvement. When encountering more extreme imbalance degree, BaCon also shows better robustness than other methods.
Authors: Roie Kazoom, Raz Birman, Ofer Hadar
Abstract: Adversarial patch attacks, crafted to compromise the integrity of Deep Neural Networks (DNNs), significantly impact Artificial Intelligence (AI) systems designed for object detection and classification tasks. The primary purpose of this work is to defend models against real-world physical attacks that target object detection and classification. We analyze attack techniques and propose a robust defense approach. We successfully reduce model confidence by over 20% using adversarial patch attacks that exploit object shape, texture and position. Leveraging the inpainting pre-processing technique, we effectively restore the original confidence levels, demonstrating the importance of robust defenses in mitigating these threats. Following fine-tuning of an AI model for traffic sign classification, we subjected it to a simulated pixelized patch-based physical adversarial attack, resulting in misclassifications. Our inpainting defense approach significantly enhances model resilience, achieving high accuracy and reliable localization despite the adversarial attacks. This contribution advances the resilience and reliability of object detection and classification networks against adversarial challenges, providing a robust foundation for critical applications.
Authors: ChungYi Lin, Shen-Lung Tung, Hung-Ting Su, Winston H. Hsu
Abstract: Vehicle flow, a crucial indicator for transportation, is often limited by detector coverage. With the advent of extensive mobile network coverage, we can leverage mobile user activities, or cellular traffic, on roadways as a proxy for vehicle flow. However, as counts of cellular traffic may not directly align with vehicle flow due to data from various user types, we present a new task: predicting vehicle flow in camera-free areas using cellular traffic. To uncover correlations within multi-source data, we deployed cameras on selected roadways to establish the Tel2Veh dataset, consisting of extensive cellular traffic and sparse vehicle flows. Addressing this challenge, we propose a framework that independently extracts features and integrates them with a graph neural network (GNN)-based fusion to discern disparities, thereby enabling the prediction of unseen vehicle flows using cellular traffic. This work advances the use of telecom data in transportation and pioneers the fusion of telecom and vision-based data, offering solutions for traffic management.
Authors: Bach Nguyen-Xuan, Thien Nguyen-Hoang, Nhu Tai-Do
Abstract: Facial Expression Recognition (FER) is a critical task within computer vision with diverse applications across various domains. Addressing the challenge of limited FER datasets, which hampers the generalization capability of expression recognition models, is imperative for enhancing performance. Our paper presents an innovative approach integrating the MAE-Face self-supervised learning (SSL) method and Fusion Attention mechanism for expression classification, particularly showcased in the 6th Affective Behavior Analysis in-the-wild (ABAW) competition. Additionally, we propose preprocessing techniques to emphasize essential facial features, thereby enhancing model performance on both training and validation sets, notably demonstrated on the Aff-wild2 dataset.
Authors: Hongyang Li, Hao Zhang, Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Lei Zhang
Abstract: In this paper, we propose a simple and strong framework for Tracking Any Point with TRansformers (TAPTR). Based on the observation that point tracking bears a great resemblance to object detection and tracking, we borrow designs from DETR-like algorithms to address the task of TAP. In the proposed framework, in each video frame, each tracking point is represented as a point query, which consists of a positional part and a content part. As in DETR, each query (its position and content feature) is naturally updated layer by layer. Its visibility is predicted by its updated content feature. Queries belonging to the same tracking point can exchange information through self-attention along the temporal dimension. As all such operations are well-designed in DETR-like algorithms, the model is conceptually very simple. We also adopt some useful designs such as cost volume from optical flow models and develop simple designs to provide long temporal information while mitigating the feature drifting issue. Our framework demonstrates strong performance with state-of-the-art performance on various TAP datasets with faster inference speed.
Authors: Baifeng Shi, Ziyang Wu, Maolin Mao, Xin Wang, Trevor Darrell
Abstract: Scaling up the size of vision models has been the de facto standard to obtain more powerful visual representations. In this work, we discuss the point beyond which larger vision models are not necessary. First, we demonstrate the power of Scaling on Scales (S$^2$), whereby a pre-trained and frozen smaller vision model (e.g., ViT-B or ViT-L), run over multiple image scales, can outperform larger models (e.g., ViT-H or ViT-G) on classification, segmentation, depth estimation, Multimodal LLM (MLLM) benchmarks, and robotic manipulation. Notably, S$^2$ achieves state-of-the-art performance in detailed understanding of MLLM on the V* benchmark, surpassing models such as GPT-4V. We examine the conditions under which S$^2$ is a preferred scaling approach compared to scaling on model size. While larger models have the advantage of better generalization on hard examples, we show that features of larger vision models can be well approximated by those of multi-scale smaller models. This suggests most, if not all, of the representations learned by current large pre-trained models can also be obtained from multi-scale smaller models. Our results show that a multi-scale smaller model has comparable learning capacity to a larger model, and pre-training smaller models with S$^2$ can match or even exceed the advantage of larger models. We release a Python package that can apply S$^2$ on any vision model with one line of code: https://github.com/bfshi/scaling_on_scales.
Authors: Hadi Alzayer, Zhihao Xia, Xuaner Zhang, Eli Shechtman, Jia-Bin Huang, Michael Gharbi
Abstract: We propose a generative model that, given a coarsely edited image, synthesizes a photorealistic output that follows the prescribed layout. Our method transfers fine details from the original image and preserves the identity of its parts. Yet, it adapts it to the lighting and context defined by the new layout. Our key insight is that videos are a powerful source of supervision for this task: objects and camera motions provide many observations of how the world changes with viewpoint, lighting, and physical interactions. We construct an image dataset in which each sample is a pair of source and target frames extracted from the same video at randomly chosen time intervals. We warp the source frame toward the target using two motion models that mimic the expected test-time user edits. We supervise our model to translate the warped image into the ground truth, starting from a pretrained diffusion model. Our model design explicitly enables fine detail transfer from the source frame to the generated image, while closely following the user-specified layout. We show that by using simple segmentations and coarse 2D manipulations, we can synthesize a photorealistic edit faithful to the user's input while addressing second-order effects like harmonizing the lighting and physical interactions between edited objects.
Authors: Armen Avetisyan, Christopher Xie, Henry Howard-Jenkins, Tsun-Yi Yang, Samir Aroudj, Suvam Patra, Fuyang Zhang, Duncan Frost, Luke Holland, Campbell Orme, Jakob Engel, Edward Miller, Richard Newcombe, Vasileios Balntas
Abstract: We introduce SceneScript, a method that directly produces full scene models as a sequence of structured language commands using an autoregressive, token-based approach. Our proposed scene representation is inspired by recent successes in transformers & LLMs, and departs from more traditional methods which commonly describe scenes as meshes, voxel grids, point clouds or radiance fields. Our method infers the set of structured language commands directly from encoded visual data using a scene language encoder-decoder architecture. To train SceneScript, we generate and release a large-scale synthetic dataset called Aria Synthetic Environments consisting of 100k high-quality in-door scenes, with photorealistic and ground-truth annotated renders of egocentric scene walkthroughs. Our method gives state-of-the art results in architectural layout estimation, and competitive results in 3D object detection. Lastly, we explore an advantage for SceneScript, which is the ability to readily adapt to new commands via simple additions to the structured language, which we illustrate for tasks such as coarse 3D object part reconstruction.
Authors: Muhammad Ridzuan, Mai Kassem, Numan Saeed, Ikboljon Sobirov, Mohammad Yaqub
Abstract: This paper introduces HuLP, a Human-in-the-Loop for Prognosis model designed to enhance the reliability and interpretability of prognostic models in clinical contexts, especially when faced with the complexities of missing covariates and outcomes. HuLP offers an innovative approach that enables human expert intervention, empowering clinicians to interact with and correct models' predictions, thus fostering collaboration between humans and AI models to produce more accurate prognosis. Additionally, HuLP addresses the challenges of missing data by utilizing neural networks and providing a tailored methodology that effectively handles missing data. Traditional methods often struggle to capture the nuanced variations within patient populations, leading to compromised prognostic predictions. HuLP imputes missing covariates based on imaging features, aligning more closely with clinician workflows and enhancing reliability. We conduct our experiments on two real-world, publicly available medical datasets to demonstrate the superiority of HuLP.
Authors: Thomas Laurent
Abstract: This paper introduces the task of "train ego-path detection", a refined approach to railway track detection designed for intelligent onboard vision systems. Whereas existing research lacks precision and often considers all tracks within the visual field uniformly, our proposed task specifically aims to identify the train's immediate path, or "ego-path", within potentially complex and dynamic railway environments. Building on this, we extend the RailSem19 dataset with ego-path annotations, facilitating further research in this direction. At the heart of our study lies TEP-Net, an end-to-end deep learning framework tailored for ego-path detection, featuring a configurable model architecture, a dynamic data augmentation strategy, and a domain-specific loss function. Leveraging a regression-based approach, TEP-Net outperforms SOTA: while addressing the track detection problem in a more nuanced way than previously, our model achieves 97.5% IoU on the test set and is faster than all existing methods. Further comparative analysis highlights the relevance of the conceptual choices behind TEP-Net, demonstrating its inherent propensity for robustness across diverse environmental conditions and operational dynamics. This work opens promising avenues for the development of intelligent driver assistance systems and autonomous train operations, paving the way toward safer and more efficient railway transportation.
Authors: Aljo\v{s}a O\v{s}ep, Tim Meinhardt, Francesco Ferroni, Neehar Peri, Deva Ramanan, Laura Leal-Taix\'e
Abstract: We propose $\texttt{SAL}$ ($\texttt{S}$egment $\texttt{A}$nything in $\texttt{L}$idar) method consisting of a text-promptable zero-shot model for segmenting and classifying any object in Lidar, and a pseudo-labeling engine that facilitates model training without manual supervision. While the established paradigm for $\textit{Lidar Panoptic Segmentation}$ (LPS) relies on manual supervision for a handful of object classes defined a priori, we utilize 2D vision foundation models to generate 3D supervision "for free". Our pseudo-labels consist of instance masks and corresponding CLIP tokens, which we lift to Lidar using calibrated multi-modal data. By training our model on these labels, we distill the 2D foundation models into our Lidar $\texttt{SAL}$ model. Even without manual labels, our model achieves $91\%$ in terms of class-agnostic segmentation and $44\%$ in terms of zero-shot LPS of the fully supervised state-of-the-art. Furthermore, we outperform several baselines that do not distill but only lift image features to 3D. More importantly, we demonstrate that $\texttt{SAL}$ supports arbitrary class prompts, can be easily extended to new datasets, and shows significant potential to improve with increasing amounts of self-labeled data.
Authors: Jurdana Masuma Iqrah, Wei Wang, Hongjie Xie, Sushil Prasad
Abstract: The observation of the advancing and retreating pattern of polar sea ice cover stands as a vital indicator of global warming. This research aims to develop a robust, effective, and scalable system for classifying polar sea ice as thick/snow-covered, young/thin, or open water using Sentinel-2 (S2) images. Since the S2 satellite is actively capturing high-resolution imagery over the earth's surface, there are lots of images that need to be classified. One major obstacle is the absence of labeled S2 training data (images) to act as the ground truth. We demonstrate a scalable and accurate method for segmenting and automatically labeling S2 images using carefully determined color thresholds. We employ a parallel workflow using PySpark to scale and achieve 9-fold data loading and 16-fold map-reduce speedup on auto-labeling S2 images based on thin cloud and shadow-filtered color-based segmentation to generate label data. The auto-labeled data generated from this process are then employed to train a U-Net machine learning model, resulting in good classification accuracy. As training the U-Net classification model is computationally heavy and time-consuming, we distribute the U-Net model training to scale it over 8 GPUs using the Horovod framework over a DGX cluster with a 7.21x speedup without affecting the accuracy of the model. Using the Antarctic's Ross Sea region as an example, the U-Net model trained on auto-labeled data achieves a classification accuracy of 98.97% for auto-labeled training datasets when the thin clouds and shadows from the S2 images are filtered out.
Authors: Hanzhou Liu, Binghan Li, Chengkai Liu, Mi Lu
Abstract: Blurry images may contain local and global non-uniform artifacts, which complicate the deblurring process and make it more challenging to achieve satisfactory results. Recently, Transformers generate improved deblurring outcomes than existing CNN architectures. However, the large model size and long inference time are still two bothersome issues which have not been fully explored. To this end, we propose DeblurDiNAT, a compact encoder-decoder Transformer which efficiently restores clean images from real-world blurry ones. We adopt an alternating dilation factor structure with the aim of global-local feature learning. Also, we observe that simply using self-attention layers in networks does not always produce good deblurred results. To solve this problem, we propose a channel modulation self-attention (CMSA) block, where a cross-channel learner (CCL) is utilized to capture channel relationships. In addition, we present a divide and multiply feed-forward network (DMFN) allowing fast feature propagation. Moreover, we design a lightweight gated feature fusion (LGFF) module, which performs controlled feature merging. Comprehensive experimental results show that the proposed model, named DeblurDiNAT, provides a favorable performance boost without introducing noticeable computational costs over the baseline, and achieves state-of-the-art (SOTA) performance on several image deblurring datasets. Compared to nearest competitors, our space-efficient and time-saving method demonstrates a stronger generalization ability with 3%-68% fewer parameters and produces deblurred images that are visually closer to the ground truth.
Authors: Yulong Shisu, Susano Mingwin, Yongshuai Wanwag, Zengqiang Chenso, Sunshin Huing
Abstract: The accurate analysis of medical images is vital for diagnosing and predicting medical conditions. Traditional approaches relying on radiologists and clinicians suffer from inconsistencies and missed diagnoses. Computer-aided diagnosis systems can assist in achieving early, accurate, and efficient diagnoses. This paper presents an improved Evolutionary Algorithm-based Transformer architecture for medical image classification using Vision Transformers. The proposed EATFormer architecture combines the strengths of Convolutional Neural Networks and Vision Transformers, leveraging their ability to identify patterns in data and adapt to specific characteristics. The architecture incorporates novel components, including the Enhanced EA-based Transformer block with Feed-Forward Network, Global and Local Interaction , and Multi-Scale Region Aggregation modules. It also introduces the Modulated Deformable MSA module for dynamic modeling of irregular locations. The paper discusses the Vision Transformer (ViT) model's key features, such as patch-based processing, positional context incorporation, and Multi-Head Attention mechanism. It introduces the Multi-Scale Region Aggregation module, which aggregates information from different receptive fields to provide an inductive bias. The Global and Local Interaction module enhances the MSA-based global module by introducing a local path for extracting discriminative local information. Experimental results on the Chest X-ray and Kvasir datasets demonstrate that the proposed EATFormer significantly improves prediction speed and accuracy compared to baseline models.
Authors: Jing Zhang, Irving Fang, Juexiao Zhang, Hao Wu, Akshat Kaushik, Alice Rodriguez, Hanwen Zhao Zhuo Zheng, Radu Iovita, Chen Feng
Abstract: Lithic Use-Wear Analysis (LUWA) using microscopic images is an underexplored vision-for-science research area. It seeks to distinguish the worked material, which is critical for understanding archaeological artifacts, material interactions, tool functionalities, and dental records. However, this challenging task goes beyond the well-studied image classification problem for common objects. It is affected by many confounders owing to the complex wear mechanism and microscopic imaging, which makes it difficult even for human experts to identify the worked material successfully. In this paper, we investigate the following three questions on this unique vision task for the first time:(i) How well can state-of-the-art pre-trained models (like DINOv2) generalize to the rarely seen domain? (ii) How can few-shot learning be exploited for scarce microscopic images? (iii) How do the ambiguous magnification and sensing modality influence the classification accuracy? To study these, we collaborated with archaeologists and built the first open-source and the largest LUWA dataset containing 23,130 microscopic images with different magnifications and sensing modalities. Extensive experiments show that existing pre-trained models notably outperform human experts but still leave a large gap for improvements. Most importantly, the LUWA dataset provides an underexplored opportunity for vision and learning communities and complements existing image classification problems on common objects.
Authors: Kasi Viswanath, Peng Jiang, Srikanth Saripalli
Abstract: LiDAR semantic segmentation frameworks predominantly leverage geometry-based features to differentiate objects within a scan. While these methods excel in scenarios with clear boundaries and distinct shapes, their performance declines in environments where boundaries are blurred, particularly in off-road contexts. To address this, recent strides in 3D segmentation algorithms have focused on harnessing raw LiDAR intensity measurements to improve prediction accuracy. Despite these efforts, current learning-based models struggle to correlate the intricate connections between raw intensity and factors such as distance, incidence angle, material reflectivity, and atmospheric conditions. Building upon our prior work, this paper delves into the advantages of employing calibrated intensity (also referred to as reflectivity) within learning-based LiDAR semantic segmentation frameworks. We initially establish that incorporating reflectivity as an input enhances the existing LiDAR semantic segmentation model. Furthermore, we present findings that enable the model to learn to calibrate intensity can boost its performance. Through extensive experimentation on the off-road dataset Rellis-3D, we demonstrate notable improvements. Specifically, converting intensity to reflectivity results in a 4% increase in mean Intersection over Union (mIoU) when compared to using raw intensity in Off-road scenarios. Additionally, we also investigate the possible benefits of using calibrated intensity in semantic segmentation in urban environments (SemanticKITTI) and cross-sensor domain adaptation.
Authors: Vincent Cartillier, Neha Jain, Irfan Essa
Abstract: We study the task of 3D multi-object re-identification from embodied tours. Specifically, an agent is given two tours of an environment (e.g. an apartment) under two different layouts (e.g. arrangements of furniture). Its task is to detect and re-identify objects in 3D - e.g. a "sofa" moved from location A to B, a new "chair" in the second layout at location C, or a "lamp" from location D in the first layout missing in the second. To support this task, we create an automated infrastructure to generate paired egocentric tours of initial/modified layouts in the Habitat simulator using Matterport3D scenes, YCB and Google-scanned objects. We present 3D Semantic MapNet (3D-SMNet) - a two-stage re-identification model consisting of (1) a 3D object detector that operates on RGB-D videos with known pose, and (2) a differentiable object matching module that solves correspondence estimation between two sets of 3D bounding boxes. Overall, 3D-SMNet builds object-based maps of each layout and then uses a differentiable matcher to re-identify objects across the tours. After training 3D-SMNet on our generated episodes, we demonstrate zero-shot transfer to real-world rearrangement scenarios by instantiating our task in Replica, Active Vision, and RIO environments depicting rearrangements. On all datasets, we find 3D-SMNet outperforms competitive baselines. Further, we show jointly training on real and generated episodes can lead to significant improvements over training on real data alone.
Authors: Konstantinos K. Delibasis, Iro Oikonomou, Aristides I. Kechriniotis, Georgios N. Tsigaridas
Abstract: A number of basic image processing tasks, such as any geometric transformation require interpolation at subpixel image values. In this work we utilize the multidimensional coordinate Hermite spline interpolation defined on non-equal spaced, rectilinear grids and apply it to a very common image processing task, image zooming. Since Hermite interpolation utilizes function values, as well as partial derivative values, it is natural to apply it to image processing tasks as a special case of equi-spaced grid, using numerical approximations of the image partial derivatives at each pixel. Furthermore, the task of image interpolation requires the calculation of image values at positions with nono-zero fractional part. Thus, any spline interpolation can be written as convolution with an appropriate kernel. In this context we generate the Hermite kernels according to the derived $n-$dimensional interpolant of Theorem 2 in [1]. We show that despite the increased complexity of the interpolant, once the kernels are constructed, the Hermite spline interpolation can be applied to images as efficiently as any other less complicated method. Finally, we perform illustrative numerical examples to showcase the applicability and high accuracy of the proposed Hermite kernels for image zooming, compared to other interpolation methods, both traditional convolution-based, as well as employing deep learning, in terms of PSNR, as well as SSIM error metrics. The proposed Hermite spline kernels outperform all other methods in the majority of the test images, in experiments using many cascaded repetitions of the zoom operation. Interesting conclusions can be drawn considering all methods under comparison.
Authors: Zaid Tasneem, Akshat Dave, Abhishek Singh, Kushagra Tiwary, Praneeth Vepakomma, Ashok Veeraraghavan, Ramesh Raskar
Abstract: Neural radiance fields (NeRFs) show potential for transforming images captured worldwide into immersive 3D visual experiences. However, most of this captured visual data remains siloed in our camera rolls as these images contain personal details. Even if made public, the problem of learning 3D representations of billions of scenes captured daily in a centralized manner is computationally intractable. Our approach, DecentNeRF, is the first attempt at decentralized, crowd-sourced NeRFs that require $\sim 10^4\times$ less server computing for a scene than a centralized approach. Instead of sending the raw data, our approach requires users to send a 3D representation, distributing the high computation cost of training centralized NeRFs between the users. It learns photorealistic scene representations by decomposing users' 3D views into personal and global NeRFs and a novel optimally weighted aggregation of only the latter. We validate the advantage of our approach to learn NeRFs with photorealism and minimal server computation cost on structured synthetic and real-world photo tourism datasets. We further analyze how secure aggregation of global NeRFs in DecentNeRF minimizes the undesired reconstruction of personal content by the server.
Authors: Anita Rau, Josiah Aklilu, F. Christopher Holsinger, Serena Yeung-Levy
Abstract: Neural Radiance Fields (NeRFs) are trained to minimize the rendering loss of predicted viewpoints. However, the photometric loss often does not provide enough information to disambiguate between different possible geometries yielding the same image. Previous work has thus incorporated depth supervision during NeRF training, leveraging dense predictions from pre-trained depth networks as pseudo-ground truth. While these depth priors are assumed to be perfect once filtered for noise, in practice, their accuracy is more challenging to capture. This work proposes a novel approach to uncertainty in depth priors for NeRF supervision. Instead of using custom-trained depth or uncertainty priors, we use off-the-shelf pretrained diffusion models to predict depth and capture uncertainty during the denoising process. Because we know that depth priors are prone to errors, we propose to supervise the ray termination distance distribution with Earth Mover's Distance instead of enforcing the rendered depth to replicate the depth prior exactly through L2-loss. Our depth-guided NeRF outperforms all baselines on standard depth metrics by a large margin while maintaining performance on photometric measures.
Authors: Austin E. Y. T. Lefebvre (Calico Life Sciences LLC), Gabriel Sturm (Calico Life Sciences LLC, Department of Biochemistry and Biophysics, University of California San Francisco), Ting-Yu Lin (Calico Life Sciences LLC), Emily Stoops (Calico Life Sciences LLC), Magdalena Preciado Lopez (Calico Life Sciences LLC), Benjamin Kaufmann-Malaga (Calico Life Sciences LLC), Kayley Hake (Calico Life Sciences LLC)
Abstract: The analysis of dynamic organelles remains a formidable challenge, though key to understanding biological processes. We introduce Nellie, an automated and unbiased pipeline for segmentation, tracking, and feature extraction of diverse intracellular structures. Nellie adapts to image metadata, eliminating user input. Nellie's preprocessing pipeline enhances structural contrast on multiple intracellular scales allowing for robust hierarchical segmentation of sub-organellar regions. Internal motion capture markers are generated and tracked via a radius-adaptive pattern matching scheme, and used as guides for sub-voxel flow interpolation. Nellie extracts a plethora of features at multiple hierarchical levels for deep and customizable analysis. Nellie features a Napari-based GUI that allows for code-free operation and visualization, while its modular open-source codebase invites customization by experienced users. We demonstrate Nellie's wide variety of use cases with two examples: unmixing multiple organelles from a single channel using feature-based classification and training an unsupervised graph autoencoder on mitochondrial multi-mesh graphs to quantify latent space embedding changes following ionomycin treatment.
Authors: Calvin Yeung, Prathyush Poduval, Mohsen Imani
Abstract: Vector Symbolic Architectures (VSAs) have emerged as a novel framework for enabling interpretable machine learning algorithms equipped with the ability to reason and explain their decision processes. The basic idea is to represent discrete information through high dimensional random vectors. Complex data structures can be built up with operations over vectors such as the "binding" operation involving element-wise vector multiplication, which associates data together. The reverse task of decomposing the associated elements is a combinatorially hard task, with an exponentially large search space. The main algorithm for performing this search is the resonator network, inspired by Hopfield network-based memory search operations. In this work, we introduce a new variant of the resonator network, based on self-attention based update rules in the iterative search problem. This update rule, based on the Hopfield network with log-sum-exp energy function and norm-bounded states, is shown to substantially improve the performance and rate of convergence. As a result, our algorithm enables a larger capacity for associative memory, enabling applications in many tasks like perception based pattern recognition, scene decomposition, and object reasoning. We substantiate our algorithm with a thorough evaluation and comparisons to baselines.
Authors: Qitong Yang, Mingtao Feng, Zijie Wu, Shijie Sun, Weisheng Dong, Yaonan Wang, Ajmal Mian
Abstract: Directly learning to model 4D content, including shape, color and motion, is challenging. Existing methods depend on skeleton-based motion control and offer limited continuity in detail. To address this, we propose a novel framework that generates coherent 4D sequences with animation of 3D shapes under given conditions with dynamic evolution of shape and color over time through integrative latent mapping. We first employ an integrative latent unified representation to encode shape and color information of each detailed 3D geometry frame. The proposed skeleton-free latent 4D sequence joint representation allows us to leverage diffusion models in a low-dimensional space to control the generation of 4D sequences. Finally, temporally coherent 4D sequences are generated conforming well to the input images and text prompts. Extensive experiments on the ShapeNet, 3DBiCar and DeformingThings4D datasets for several tasks demonstrate that our method effectively learns to generate quality 3D shapes with color and 4D mesh animations, improving over the current state-of-the-art. Source code will be released.
Authors: Zhengqing Yuan, Ruoxi Chen, Zhaoxu Li, Haolong Jia, Lifang He, Chi Wang, Lichao Sun
Abstract: Sora is the first large-scale generalist video generation model that garnered significant attention across society. Since its launch by OpenAI in February 2024, no other video generation models have paralleled {Sora}'s performance or its capacity to support a broad spectrum of video generation tasks. Additionally, there are only a few fully published video generation models, with the majority being closed-source. To address this gap, this paper proposes a new multi-agent framework Mora, which incorporates several advanced visual AI agents to replicate generalist video generation demonstrated by Sora. In particular, Mora can utilize multiple visual agents and successfully mimic Sora's video generation capabilities in various tasks, such as (1) text-to-video generation, (2) text-conditional image-to-video generation, (3) extend generated videos, (4) video-to-video editing, (5) connect videos and (6) simulate digital worlds. Our extensive experimental results show that Mora achieves performance that is proximate to that of Sora in various tasks. However, there exists an obvious performance gap between our work and Sora when assessed holistically. In summary, we hope this project can guide the future trajectory of video generation through collaborative AI agents.
Authors: Xian Lin, Yangyang Xiang, Zhehao Wang, Kwang-Ting Cheng, Zengqiang Yan, Li Yu
Abstract: Segment anything model (SAM), a foundation model with superior versatility and generalization across diverse segmentation tasks, has attracted widespread attention in medical imaging. However, it has been proved that SAM would encounter severe performance degradation due to the lack of medical knowledge in training and local feature encoding. Though several SAM-based models have been proposed for tuning SAM in medical imaging, they still suffer from insufficient feature extraction and highly rely on high-quality prompts. In this paper, we construct a large CT dataset consisting of 1.1M CT images and 5M masks from public datasets and propose a powerful foundation model SAMCT allowing labor-free prompts. Specifically, based on SAM, SAMCT is further equipped with a U-shaped CNN image encoder, a cross-branch interaction module, and a task-indicator prompt encoder. The U-shaped CNN image encoder works in parallel with the ViT image encoder in SAM to supplement local features. Cross-branch interaction enhances the feature expression capability of the CNN image encoder and the ViT image encoder by exchanging global perception and local features from one to the other. The task-indicator prompt encoder is a plug-and-play component to effortlessly encode task-related indicators into prompt embeddings. In this way, SAMCT can work in an automatic manner in addition to the semi-automatic interactive strategy in SAM. Extensive experiments demonstrate the superiority of SAMCT against the state-of-the-art task-specific and SAM-based medical foundation models on various tasks. The code, data, and models are released at https://github.com/xianlin7/SAMCT.
Authors: Kewei Wang, Yizheng Wu, Jun Cen, Zhiyu Pan, Xingyi Li, Zhe Wang, Zhiguo Cao, Guosheng Lin
Abstract: The perception of motion behavior in a dynamic environment holds significant importance for autonomous driving systems, wherein class-agnostic motion prediction methods directly predict the motion of the entire point cloud. While most existing methods rely on fully-supervised learning, the manual labeling of point cloud data is laborious and time-consuming. Therefore, several annotation-efficient methods have been proposed to address this challenge. Although effective, these methods rely on weak annotations or additional multi-modal data like images, and the potential benefits inherent in the point cloud sequence are still underexplored. To this end, we explore the feasibility of self-supervised motion prediction with only unlabeled LiDAR point clouds. Initially, we employ an optimal transport solver to establish coarse correspondences between current and future point clouds as the coarse pseudo motion labels. Training models directly using such coarse labels leads to noticeable spatial and temporal prediction inconsistencies. To mitigate these issues, we introduce three simple spatial and temporal regularization losses, which facilitate the self-supervised training process effectively. Experimental results demonstrate the significant superiority of our approach over the state-of-the-art self-supervised methods.
Authors: Tongtian Yue, Jie Cheng, Longteng Guo, Xingyuan Dai, Zijia Zhao, Xingjian He, Gang Xiong, Yisheng Lv, Jing Liu
Abstract: Recent trends in Large Vision Language Models (LVLMs) research have been increasingly focusing on advancing beyond general image understanding towards more nuanced, object-level referential comprehension. In this paper, we present and delve into the self-consistency capability of LVLMs, a crucial aspect that reflects the models' ability to both generate informative captions for specific objects and subsequently utilize these captions to accurately re-identify the objects in a closed-loop process. This capability significantly mirrors the precision and reliability of fine-grained visual-language understanding. Our findings reveal that the self-consistency level of existing LVLMs falls short of expectations, posing limitations on their practical applicability and potential. To address this gap, we introduce a novel fine-tuning paradigm named Self-Consistency Tuning (SC-Tune). It features the synergistic learning of a cyclic describer-locator system. This paradigm is not only data-efficient but also exhibits generalizability across multiple LVLMs. Through extensive experiments, we demonstrate that SC-Tune significantly elevates performance across a spectrum of object-level vision-language benchmarks and maintains competitive or improved performance on image-level vision-language benchmarks. Both our model and code will be publicly available at https://github.com/ivattyue/SC-Tune.
Authors: Mengyu Yang, Ye Tian, Lanshan Zhang, Xiao Liang, Xuming Ran, Wendong Wang
Abstract: Recently, prompt-based methods have emerged as a new alternative `parameter-efficient fine-tuning' paradigm, which only fine-tunes a small number of additional parameters while keeping the original model frozen. However, despite achieving notable results, existing prompt methods mainly focus on `what to add', while overlooking the equally important aspect of `where to add', typically relying on the manually crafted placement. To this end, we propose a region-based Adaptive Visual Prompt, named AdaViPro, which integrates the `where to add' optimization of the prompt into the learning process. Specifically, we reconceptualize the `where to add' optimization as a problem of regional decision-making. During inference, AdaViPro generates a regionalized mask map for the whole image, which is composed of 0 and 1, to designate whether to apply or discard the prompt in each specific area. Therefore, we employ Gumbel-Softmax sampling to enable AdaViPro's end-to-end learning through standard back-propagation. Extensive experiments demonstrate that our AdaViPro yields new efficiency and accuracy trade-offs for adapting pre-trained models.
Authors: Han-Hung Lee, Manolis Savva, Angel X. Chang
Abstract: Recent years have seen an explosion of work and interest in text-to-3D shape generation. Much of the progress is driven by advances in 3D representations, large-scale pretraining and representation learning for text and image data enabling generative AI models, and differentiable rendering. Computational systems that can perform text-to-3D shape generation have captivated the popular imagination as they enable non-expert users to easily create 3D content directly from text. However, there are still many limitations and challenges remaining in this problem space. In this state-of-the-art report, we provide a survey of the underlying technology and methods enabling text-to-3D shape generation to summarize the background literature. We then derive a systematic categorization of recent work on text-to-3D shape generation based on the type of supervision data required. Finally, we discuss limitations of the existing categories of methods, and delineate promising directions for future work.
Authors: Keith G. Mills, Fred X. Han, Mohammad Salameh, Shengyao Lu, Chunhua Zhou, Jiao He, Fengyu Sun, Di Niu
Abstract: Neural Architecture Search is a costly practice. The fact that a search space can span a vast number of design choices with each architecture evaluation taking nontrivial overhead makes it hard for an algorithm to sufficiently explore candidate networks. In this paper, we propose AutoBuild, a scheme which learns to align the latent embeddings of operations and architecture modules with the ground-truth performance of the architectures they appear in. By doing so, AutoBuild is capable of assigning interpretable importance scores to architecture modules, such as individual operation features and larger macro operation sequences such that high-performance neural networks can be constructed without any need for search. Through experiments performed on state-of-the-art image classification, segmentation, and Stable Diffusion models, we show that by mining a relatively small set of evaluated architectures, AutoBuild can learn to build high-quality architectures directly or help to reduce search space to focus on relevant areas, finding better architectures that outperform both the original labeled ones and ones found by search baselines. Code available at https://github.com/Ascend-Research/AutoBuild
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: Yibo Wang, Ruiyuan Gao, Kai Chen, Kaiqiang Zhou, Yingjie Cai, Lanqing Hong, Zhenguo Li, Lihui Jiang, Dit-Yan Yeung, Qiang Xu, Kai Zhang
Abstract: Current perceptive models heavily depend on resource-intensive datasets, prompting the need for innovative solutions. Leveraging recent advances in diffusion models, synthetic data, by constructing image inputs from various annotations, proves beneficial for downstream tasks. While prior methods have separately addressed generative and perceptive models, DetDiffusion, for the first time, harmonizes both, tackling the challenges in generating effective data for perceptive models. To enhance image generation with perceptive models, we introduce perception-aware loss (P.A. loss) through segmentation, improving both quality and controllability. To boost the performance of specific perceptive models, our method customizes data augmentation by extracting and utilizing perception-aware attribute (P.A. Attr) during generation. Experimental results from the object detection task highlight DetDiffusion's superior performance, establishing a new state-of-the-art in layout-guided generation. Furthermore, image syntheses from DetDiffusion can effectively augment training data, significantly enhancing downstream detection performance.
Authors: Peishan Cong, Ziyi WangZhiyang Dou, Yiming Ren, Wei Yin, Kai Cheng, Yujing Sun, Xiaoxiao Long, Xinge Zhu, Yuexin Ma
Abstract: Language-guided scene-aware human motion generation has great significance for entertainment and robotics. In response to the limitations of existing datasets, we introduce LaserHuman, a pioneering dataset engineered to revolutionize Scene-Text-to-Motion research. LaserHuman stands out with its inclusion of genuine human motions within 3D environments, unbounded free-form natural language descriptions, a blend of indoor and outdoor scenarios, and dynamic, ever-changing scenes. Diverse modalities of capture data and rich annotations present great opportunities for the research of conditional motion generation, and can also facilitate the development of real-life applications. Moreover, to generate semantically consistent and physically plausible human motions, we propose a multi-conditional diffusion model, which is simple but effective, achieving state-of-the-art performance on existing datasets.
Authors: Yew Ken Chia, Vernon Toh Yan Han, Deepanway Ghosal, Lidong Bing, Soujanya Poria
Abstract: Large multimodal models extend the impressive capabilities of large language models by integrating multimodal understanding abilities. However, it is not clear how they can emulate the general intelligence and reasoning ability of humans. As recognizing patterns and abstracting concepts are key to general intelligence, we introduce PuzzleVQA, a collection of puzzles based on abstract patterns. With this dataset, we evaluate large multimodal models with abstract patterns based on fundamental concepts, including colors, numbers, sizes, and shapes. Through our experiments on state-of-the-art large multimodal models, we find that they are not able to generalize well to simple abstract patterns. Notably, even GPT-4V cannot solve more than half of the puzzles. To diagnose the reasoning challenges in large multimodal models, we progressively guide the models with our ground truth reasoning explanations for visual perception, inductive reasoning, and deductive reasoning. Our systematic analysis finds that the main bottlenecks of GPT-4V are weaker visual perception and inductive reasoning abilities. Through this work, we hope to shed light on the limitations of large multimodal models and how they can better emulate human cognitive processes in the future (Our data and code will be released publicly at https://github.com/declare-lab/LLM-PuzzleTest).
Authors: Daniel Duenias, Brennan Nichyporuk, Tal Arbel, Tammy Riklin Raviv
Abstract: The integration of diverse clinical modalities such as medical imaging and the tabular data obtained by the patients' Electronic Health Records (EHRs) is a crucial aspect of modern healthcare. The integrative analysis of multiple sources can provide a comprehensive understanding of a patient's condition and can enhance diagnoses and treatment decisions. Deep Neural Networks (DNNs) consistently showcase outstanding performance in a wide range of multimodal tasks in the medical domain. However, the complex endeavor of effectively merging medical imaging with clinical, demographic and genetic information represented as numerical tabular data remains a highly active and ongoing research pursuit. We present a novel framework based on hypernetworks to fuse clinical imaging and tabular data by conditioning the image processing on the EHR's values and measurements. This approach aims to leverage the complementary information present in these modalities to enhance the accuracy of various medical applications. We demonstrate the strength and the generality of our method on two different brain Magnetic Resonance Imaging (MRI) analysis tasks, namely, brain age prediction conditioned by subject's sex, and multiclass Alzheimer's Disease (AD) classification conditioned by tabular data. We show that our framework outperforms both single-modality models and state-of-the-art MRI-tabular data fusion methods. The code, enclosed to this manuscript will be made publicly available.
Authors: Yifan Wu, Jiawei Du, Ping Liu, Yuewei Lin, Wenqing Cheng, Wei Xu
Abstract: Dataset distillation is an advanced technique aimed at compressing datasets into significantly smaller counterparts, while preserving formidable training performance. Significant efforts have been devoted to promote evaluation accuracy under limited compression ratio while overlooked the robustness of distilled dataset. In this work, we introduce a comprehensive benchmark that, to the best of our knowledge, is the most extensive to date for evaluating the adversarial robustness of distilled datasets in a unified way. Our benchmark significantly expands upon prior efforts by incorporating a wider range of dataset distillation methods, including the latest advancements such as TESLA and SRe2L, a diverse array of adversarial attack methods, and evaluations across a broader and more extensive collection of datasets such as ImageNet-1K. Moreover, we assessed the robustness of these distilled datasets against representative adversarial attack algorithms like PGD and AutoAttack, while exploring their resilience from a frequency perspective. We also discovered that incorporating distilled data into the training batches of the original dataset can yield to improvement of robustness.
Authors: K Huang, G Song, Hanwen Su, Jiyan Wang
Abstract: Out-of-distribution (OOD) detection is a critical task to ensure the reliability and security of machine learning models deployed in real-world applications. Conventional methods for OOD detection that rely on single-modal information, often struggle to capture the rich variety of OOD instances. The primary difficulty in OOD detection arises when an input image has numerous similarities to a particular class in the in-distribution (ID) dataset, e.g., wolf to dog, causing the model to misclassify it. Nevertheless, it may be easy to distinguish these classes in the semantic domain. To this end, in this paper, a novel method called ODPC is proposed, in which specific prompts to generate OOD peer classes of ID semantics are designed by a large language model as an auxiliary modality to facilitate detection. Moreover, a contrastive loss based on OOD peer classes is devised to learn compact representations of ID classes and improve the clarity of boundaries between different classes. The extensive experiments on five benchmark datasets show that the method we propose can yield state-of-the-art results.
Authors: Otto Seiskari, Jerry Ylilammi, Valtteri Kaatrasalo, Pekka Rantalankila, Matias Turkulainen, Juho Kannala, Esa Rahtu, Arno Solin
Abstract: High-quality scene reconstruction and novel view synthesis based on Gaussian Splatting (3DGS) typically require steady, high-quality photographs, often impractical to capture with handheld cameras. We present a method that adapts to camera motion and allows high-quality scene reconstruction with handheld video data suffering from motion blur and rolling shutter distortion. Our approach is based on detailed modelling of the physical image formation process and utilizes velocities estimated using visual-inertial odometry (VIO). Camera poses are considered non-static during the exposure time of a single image frame and camera poses are further optimized in the reconstruction process. We formulate a differentiable rendering pipeline that leverages screen space approximation to efficiently incorporate rolling-shutter and motion blur effects into the 3DGS framework. Our results with both synthetic and real data demonstrate superior performance in mitigating camera motion over existing methods, thereby advancing 3DGS in naturalistic settings.
Authors: LeoWu TomyEnrique, Xiangcheng Du, Kangliang Liu, Han Yuan, Zhao Zhou, Cheng Jin
Abstract: Scene text image super-resolution has significantly improved the accuracy of scene text recognition. However, many existing methods emphasize performance over efficiency and ignore the practical need for lightweight solutions in deployment scenarios. Faced with the issues, our work proposes an efficient framework called SGENet to facilitate deployment on resource-limited platforms. SGENet contains two branches: super-resolution branch and semantic guidance branch. We apply a lightweight pre-trained recognizer as a semantic extractor to enhance the understanding of text information. Meanwhile, we design the visual-semantic alignment module to achieve bidirectional alignment between image features and semantics, resulting in the generation of highquality prior guidance. We conduct extensive experiments on benchmark dataset, and the proposed SGENet achieves excellent performance with fewer computational costs. Code is available at https://github.com/SijieLiu518/SGENet
Authors: Xiaosong Jia, Shaoshuai Shi, Zijun Chen, Li Jiang, Wenlong Liao, Tao He, Junchi Yan
Abstract: As an essential task in autonomous driving (AD), motion prediction aims to predict the future states of surround objects for navigation. One natural solution is to estimate the position of other agents in a step-by-step manner where each predicted time-step is conditioned on both observed time-steps and previously predicted time-steps, i.e., autoregressive prediction. Pioneering works like SocialLSTM and MFP design their decoders based on this intuition. However, almost all state-of-the-art works assume that all predicted time-steps are independent conditioned on observed time-steps, where they use a single linear layer to generate positions of all time-steps simultaneously. They dominate most motion prediction leaderboards due to the simplicity of training MLPs compared to autoregressive networks. In this paper, we introduce the GPT style next token prediction into motion forecasting. In this way, the input and output could be represented in a unified space and thus the autoregressive prediction becomes more feasible. However, different from language data which is composed of homogeneous units -words, the elements in the driving scene could have complex spatial-temporal and semantic relations. To this end, we propose to adopt three factorized attention modules with different neighbors for information aggregation and different position encoding styles to capture their relations, e.g., encoding the transformation between coordinate systems for spatial relativity while adopting RoPE for temporal relativity. Empirically, by equipping with the aforementioned tailored designs, the proposed method achieves state-of-the-art performance in the Waymo Open Motion and Waymo Interaction datasets. Notably, AMP outperforms other recent autoregressive motion prediction methods: MotionLM and StateTransformer, which demonstrates the effectiveness of the proposed designs.
Authors: Quan Zheng, Hao Sun, Huiyao Xu, Fanjiang Xu
Abstract: Neural radiance field has achieved fundamental success in novel view synthesis from input views with the same brightness level captured under fixed normal lighting. Unfortunately, synthesizing novel views remains to be a challenge for input views with heterogeneous brightness level captured under low-light condition. The condition is pretty common in the real world. It causes low-contrast images where details are concealed in the darkness and camera sensor noise significantly degrades the image quality. To tackle this problem, we propose to learn to decompose illumination, reflectance, and noise from input views according to that reflectance remains invariant across heterogeneous views. To cope with heterogeneous brightness and noise levels across multi-views, we learn an illumination embedding and optimize a noise map individually for each view. To allow intuitive editing of the illumination, we design an illumination adjustment module to enable either brightening or darkening of the illumination component. Comprehensive experiments demonstrate that this approach enables effective intrinsic decomposition for low-light multi-view noisy images and achieves superior visual quality and numerical performance for synthesizing novel views compared to state-of-the-art methods.
Authors: Yilin Leng (for the Alzheimer's Disease Neuroimaging Initiative), Wenju Cui (for the Alzheimer's Disease Neuroimaging Initiative), Bai Chen (for the Alzheimer's Disease Neuroimaging Initiative), Xi Jiang (for the Alzheimer's Disease Neuroimaging Initiative), Shuangqing Chen (for the Alzheimer's Disease Neuroimaging Initiative), Jian Zheng (for the Alzheimer's Disease Neuroimaging Initiative)
Abstract: Prediction the conversion to early-stage dementia is critical for mitigating its progression but remains challenging due to subtle cognitive impairments and structural brain changes. Traditional T1-weighted magnetic resonance imaging (T1-MRI) research focus on identifying brain atrophy regions but often fails to address the intricate connectivity between them. This limitation underscores the necessity of focuing on inter-regional connectivity for a comprehensive understand of the brain's complex network. Moreover, there is a pressing demand for methods that adaptively preserve and extract critical information, particularly specialized subgraph mining techniques for brain networks. These are essential for developing high-quality feature representations that reveal critical spatial impacts of structural brain changes and its topology. In this paper, we propose Brain-SubGNN, a novel graph representation network to mine and enhance critical subgraphs based on T1-MRI. This network provides a subgraph-level interpretation, enhancing interpretability and insights for graph analysis. The process begins by extracting node features and a correlation matrix between nodes to construct a task-oriented brain network. Brain-SubGNN then adaptively identifies and enhances critical subgraphs, capturing both loop and neighbor subgraphs. This method reflects the loop topology and local changes, indicative of long-range connections, and maintains local and global brain attributes. Extensive experiments validate the effectiveness and advantages of Brain-SubGNN, demonstrating its potential as a powerful tool for understanding and diagnosing early-stage dementia. Source code is available at https://github.com/Leng-10/Brain-SubGNN.
Authors: Santosh Sanjeev, Nuren Zhaksylyk, Ibrahim Almakky, Anees Ur Rehman Hashmi, Mohammad Areeb Qazi, Mohammad Yaqub
Abstract: The scarcity of well-annotated medical datasets requires leveraging transfer learning from broader datasets like ImageNet or pre-trained models like CLIP. Model soups averages multiple fine-tuned models aiming to improve performance on In-Domain (ID) tasks and enhance robustness against Out-of-Distribution (OOD) datasets. However, applying these methods to the medical imaging domain faces challenges and results in suboptimal performance. This is primarily due to differences in error surface characteristics that stem from data complexities such as heterogeneity, domain shift, class imbalance, and distributional shifts between training and testing phases. To address this issue, we propose a hierarchical merging approach that involves local and global aggregation of models at various levels based on models' hyperparameter configurations. Furthermore, to alleviate the need for training a large number of models in the hyperparameter search, we introduce a computationally efficient method using a cyclical learning rate scheduler to produce multiple models for aggregation in the weight space. Our method demonstrates significant improvements over the model souping approach across multiple datasets (around 6% gain in HAM10000 and CheXpert datasets) while maintaining low computational costs for model generation and selection. Moreover, we achieve better results on OOD datasets than model soups. The code is available at https://github.com/BioMedIA-MBZUAI/FissionFusion.
Authors: Santosh Sanjeev, Fadillah Adamsyah Maani, Arsen Abzhanov, Vijay Ram Papineni, Ibrahim Almakky, Bart{\l}omiej W. Papie\.z, Mohammad Yaqub
Abstract: With the emergence of vision language models in the medical imaging domain, numerous studies have focused on two dominant research activities: (1) report generation from Chest X-rays (CXR), and (2) synthetic scan generation from text or reports. Despite some research incorporating multi-view CXRs into the generative process, prior patient scans and reports have been generally disregarded. This can inadvertently lead to the leaving out of important medical information, thus affecting generation quality. To address this, we propose TiBiX: Leveraging Temporal information for Bidirectional X-ray and Report Generation. Considering previous scans, our approach facilitates bidirectional generation, primarily addressing two challenging problems: (1) generating the current image from the previous image and current report and (2) generating the current report based on both the previous and current images. Moreover, we extract and release a curated temporal benchmark dataset derived from the MIMIC-CXR dataset, which focuses on temporal data. Our comprehensive experiments and ablation studies explore the merits of incorporating prior CXRs and achieve state-of-the-art (SOTA) results on the report generation task. Furthermore, we attain on-par performance with SOTA image generation efforts, thus serving as a new baseline in longitudinal bidirectional CXR-to-report generation. The code is available at https://github.com/BioMedIA-MBZUAI/TiBiX.
Authors: Joonmyung Choi, Sanghyeok Lee, Jaewon Chu, Minhyuk Choi, Hyunwoo J. Kim
Abstract: Video Transformers have become the prevalent solution for various video downstream tasks with superior expressive power and flexibility. However, these video transformers suffer from heavy computational costs induced by the massive number of tokens across the entire video frames, which has been the major barrier to training the model. Further, the patches irrelevant to the main contents, e.g., backgrounds, degrade the generalization performance of models. To tackle these issues, we propose training free token merging for lightweight video Transformer (vid-TLDR) that aims to enhance the efficiency of video Transformers by merging the background tokens without additional training. For vid-TLDR, we introduce a novel approach to capture the salient regions in videos only with the attention map. Further, we introduce the saliency-aware token merging strategy by dropping the background tokens and sharpening the object scores. Our experiments show that vid-TLDR significantly mitigates the computational complexity of video Transformers while achieving competitive performance compared to the base model without vid-TLDR. Code is available at https://github.com/mlvlab/vid-TLDR.
Authors: Xinyu Geng, Jiaming Wang, Jiawei Gong, Yuerong Xue, Jun Xu, Fanglin Chen, Xiaolin Huang
Abstract: Redundancy is a persistent challenge in Capsule Networks (CapsNet),leading to high computational costs and parameter counts. Although previous works have introduced pruning after the initial capsule layer, dynamic routing's fully connected nature and non-orthogonal weight matrices reintroduce redundancy in deeper layers. Besides, dynamic routing requires iterating to converge, further increasing computational demands. In this paper, we propose an Orthogonal Capsule Network (OrthCaps) to reduce redundancy, improve routing performance and decrease parameter counts. Firstly, an efficient pruned capsule layer is introduced to discard redundant capsules. Secondly, dynamic routing is replaced with orthogonal sparse attention routing, eliminating the need for iterations and fully connected structures. Lastly, weight matrices during routing are orthogonalized to sustain low capsule similarity, which is the first approach to introduce orthogonality into CapsNet as far as we know. Our experiments on baseline datasets affirm the efficiency and robustness of OrthCaps in classification tasks, in which ablation studies validate the criticality of each component. Remarkably, OrthCaps-Shallow outperforms other Capsule Network benchmarks on four datasets, utilizing only 110k parameters, which is a mere 1.25% of a standard Capsule Network's total. To the best of our knowledge, it achieves the smallest parameter count among existing Capsule Networks. Similarly, OrthCaps-Deep demonstrates competitive performance across four datasets, utilizing only 1.2% of the parameters required by its counterparts.
Authors: Jingkun An, Yinghao Zhu, Zongjian Li, Haoran Feng, Bohua Chen, Yemin Shi, Chengwei Pan
Abstract: Text-to-Image (T2I) diffusion models have achieved remarkable success in image generation. Despite their progress, challenges remain in both prompt-following ability, image quality and lack of high-quality datasets, which are essential for refining these models. As acquiring labeled data is costly, we introduce AGFSync, a framework that enhances T2I diffusion models through Direct Preference Optimization (DPO) in a fully AI-driven approach. AGFSync utilizes Vision-Language Models (VLM) to assess image quality across style, coherence, and aesthetics, generating feedback data within an AI-driven loop. By applying AGFSync to leading T2I models such as SD v1.4, v1.5, and SDXL, our extensive experiments on the TIFA dataset demonstrate notable improvements in VQA scores, aesthetic evaluations, and performance on the HPSv2 benchmark, consistently outperforming the base models. AGFSync's method of refining T2I diffusion models paves the way for scalable alignment techniques.
Authors: Kaito Shiku, Shinnosuke Matsuo, Daiki Suehiro, Ryoma Bise
Abstract: The paper proposes a novel problem in multi-class Multiple-Instance Learning (MIL) called Learning from the Majority Label (LML). In LML, the majority class of instances in a bag is assigned as the bag's label. LML aims to classify instances using bag-level majority classes. This problem is valuable in various applications. Existing MIL methods are unsuitable for LML due to aggregating confidences, which may lead to inconsistency between the bag-level label and the label obtained by counting the number of instances for each class. This may lead to incorrect instance-level classification. We propose a counting network trained to produce the bag-level majority labels estimated by counting the number of instances for each class. This led to the consistency of the majority class between the network outputs and one obtained by counting the number of instances. Experimental results show that our counting network outperforms conventional MIL methods on four datasets The code is publicly available at https://github.com/Shiku-Kaito/Counting-Network-for-Learning-from-Majority-Label.
URLs: https://github.com/Shiku-Kaito/Counting-Network-for-Learning-from-Majority-Label.
Authors: Jiawei Zhou, Wuzhou Li, Yi Cao, Hongtao Cai, Xiang Li
Abstract: Few-shot object detection (FSOD) has garnered significant research attention in the field of remote sensing due to its ability to reduce the dependency on large amounts of annotated data. However, two challenges persist in this area: (1) axis-aligned proposals, which can result in misalignment for arbitrarily oriented objects, and (2) the scarcity of annotated data still limits the performance for unseen object categories. To address these issues, we propose a novel FSOD method for remote sensing images called Few-shot Oriented object detection with Memorable Contrastive learning (FOMC). Specifically, we employ oriented bounding boxes instead of traditional horizontal bounding boxes to learn a better feature representation for arbitrary-oriented aerial objects, leading to enhanced detection performance. To the best of our knowledge, we are the first to address oriented object detection in the few-shot setting for remote sensing images. To address the challenging issue of object misclassification, we introduce a supervised contrastive learning module with a dynamically updated memory bank. This module enables the use of large batches of negative samples and enhances the model's capability to learn discriminative features for unseen classes. We conduct comprehensive experiments on the DOTA and HRSC2016 datasets, and our model achieves state-of-the-art performance on the few-shot oriented object detection task. Code and pretrained models will be released.
Authors: Jannik Presberger, Rashmiparvathi Keshara, David Stein, Yung Hae Kim, Anne Grapin-Botton, Bjoern Andres
Abstract: In biological and medical research, scientists now routinely acquire microscopy images of hundreds of morphologically heterogeneous organoids and are then faced with the task of finding patterns in the image collection, i.e., subsets of organoids that appear similar and potentially represent the same morphological class. We adopt models and algorithms for correlating organoid images, i.e., for quantifying the similarity in appearance and geometry of the organoids they depict, and for clustering organoid images by consolidating conflicting correlations. For correlating organoid images, we adopt and compare two alternatives, a partial quadratic assignment problem and a twin network. For clustering organoid images, we employ the correlation clustering problem. Empirically, we learn the parameters of these models, infer a clustering of organoid images, and quantify the accuracy of the inferred clusters, with respect to a training set and a test set we contribute of state-of-the-art light microscopy images of organoids clustered manually by biologists.
Authors: Feng Liu, Xiaobin-Chang
Abstract: Semantic image synthesis aims to generate high-quality images given semantic conditions, i.e. segmentation masks and style reference images. Existing methods widely adopt generative adversarial networks (GANs). GANs take all conditional inputs and directly synthesize images in a single forward step. In this paper, semantic image synthesis is treated as an image denoising task and is handled with a novel image-to-image diffusion model (IIDM). Specifically, the style reference is first contaminated with random noise and then progressively denoised by IIDM, guided by segmentation masks. Moreover, three techniques, refinement, color-transfer and model ensembles, are proposed to further boost the generation quality. They are plug-in inference modules and do not require additional training. Extensive experiments show that our IIDM outperforms existing state-of-the-art methods by clear margins. Further analysis is provided via detailed demonstrations. We have implemented IIDM based on the Jittor framework; code is available at https://github.com/ader47/jittor-jieke-semantic_images_synthesis.
URLs: https://github.com/ader47/jittor-jieke-semantic_images_synthesis.
Authors: Wenqi Zhao
Abstract: In order to improve the robustness of traditional image segmentation models to noise, this paper models the illumination term in intensity inhomogeneity images. Additionally, to enhance the model's robustness to noisy images, we incorporate the binary level set model into the proposed model. Compared to the traditional level set, the binary level set eliminates the need for continuous reinitialization. Moreover, by introducing the variational operator GL, our model demonstrates better capability in segmenting noisy images. Finally, we employ the three-step splitting operator method for solving, and the effectiveness of the proposed model is demonstrated on various images.
Authors: Alberto Garc\'ia-Hern\'andez, Riccardo Giubilato, Klaus H. Strobl, Javier Civera, Rudolph Triebel
Abstract: Perceptual aliasing and weak textures pose significant challenges to the task of place recognition, hindering the performance of Simultaneous Localization and Mapping (SLAM) systems. This paper presents a novel model, called UMF (standing for Unifying Local and Global Multimodal Features) that 1) leverages multi-modality by cross-attention blocks between vision and LiDAR features, and 2) includes a re-ranking stage that re-orders based on local feature matching the top-k candidates retrieved using a global representation. Our experiments, particularly on sequences captured on a planetary-analogous environment, show that UMF outperforms significantly previous baselines in those challenging aliased environments. Since our work aims to enhance the reliability of SLAM in all situations, we also explore its performance on the widely used RobotCar dataset, for broader applicability. Code and models are available at https://github.com/DLR-RM/UMF
Authors: Yamin Mao, Zhihua Liu, Weiming Li, SoonYong Cho, Qiang Wang, Xiaoshuai Hao
Abstract: Depth-based 3D hand pose estimation is an important but challenging research task in human-machine interaction community. Recently, dense regression methods have attracted increasing attention in 3D hand pose estimation task, which provide a low computational burden and high accuracy regression way by densely regressing hand joint offset maps. However, large-scale regression offset values are often affected by noise and outliers, leading to a significant drop in accuracy. To tackle this, we re-formulate 3D hand pose estimation as a dense ordinal regression problem and propose a novel Dense Ordinal Regression 3D Pose Network (DOR3D-Net). Specifically, we first decompose offset value regression into sub-tasks of binary classifications with ordinal constraints. Then, each binary classifier can predict the probability of a binary spatial relationship relative to joint, which is easier to train and yield much lower level of noise. The estimated hand joint positions are inferred by aggregating the ordinal regression results at local positions with a weighted sum. Furthermore, both joint regression loss and ordinal regression loss are used to train our DOR3D-Net in an end-to-end manner. Extensive experiments on public datasets (ICVL, MSRA, NYU and HANDS2017) show that our design provides significant improvements over SOTA methods.
Authors: Haoran Lang, Yuxuan Ge, Zheng Tian
Abstract: Diffusion models have achieved great success in image generation. However, when leveraging this idea for video generation, we face significant challenges in maintaining the consistency and continuity across video frames. This is mainly caused by the lack of an effective framework to align frames of videos with desired temporal features while preserving consistent semantic and stochastic features. In this work, we propose a novel Sector-Shaped Diffusion Model (S2DM) whose sector-shaped diffusion region is formed by a set of ray-shaped reverse diffusion processes starting at the same noise point. S2DM can generate a group of intrinsically related data sharing the same semantic and stochastic features while varying on temporal features with appropriate guided conditions. We apply S2DM to video generation tasks, and explore the use of optical flow as temporal conditions. Our experimental results show that S2DM outperforms many existing methods in the task of video generation without any temporal-feature modelling modules. For text-to-video generation tasks where temporal conditions are not explicitly given, we propose a two-stage generation strategy which can decouple the generation of temporal features from semantic-content features. We show that, without additional training, our model integrated with another temporal conditions generative model can still achieve comparable performance with existing works. Our results can be viewd at https://s2dm.github.io/S2DM/.
Authors: Kaito Shiku, Hiromitsu Shirai, Takeshi Ishihara, Ryoma Bise
Abstract: 3D cell tracking in a living organism has a crucial role in live cell image analysis. Cell tracking in C. elegans has two difficulties. First, cell migration in a consecutive frame is large since they move their head during scanning. Second, cell detection is often inconsistent in consecutive frames due to touching cells and low-contrast images, and these inconsistent detections affect the tracking performance worse. In this paper, we propose a cell tracking method to address these issues, which has two main contributions. First, we introduce cell position heatmap-based non-rigid alignment with test-time fine-tuning, which can warp the detected points to near the positions at the next frame. Second, we propose a pairwise detection method, which uses the information of detection results at the previous frame for detecting cells at the current frame. The experimental results demonstrate the effectiveness of each module, and the proposed method achieved the best performance in comparison.
Authors: Yicheng Wu, Xiangde Luo, Zhe Xu, Xiaoqing Guo, Lie Ju, Zongyuan Ge, Wenjun Liao, Jianfei Cai
Abstract: Annotation ambiguity due to inherent data uncertainties such as blurred boundaries in medical scans and different observer expertise and preferences has become a major obstacle for training deep-learning based medical image segmentation models. To address it, the common practice is to gather multiple annotations from different experts, leading to the setting of multi-rater medical image segmentation. Existing works aim to either merge different annotations into the "groundtruth" that is often unattainable in numerous medical contexts, or generate diverse results, or produce personalized results corresponding to individual expert raters. Here, we bring up a more ambitious goal for multi-rater medical image segmentation, i.e., obtaining both diversified and personalized results. Specifically, we propose a two-stage framework named D-Persona (first Diversification and then Personalization). In Stage I, we exploit multiple given annotations to train a Probabilistic U-Net model, with a bound-constrained loss to improve the prediction diversity. In this way, a common latent space is constructed in Stage I, where different latent codes denote diversified expert opinions. Then, in Stage II, we design multiple attention-based projection heads to adaptively query the corresponding expert prompts from the shared latent space, and then perform the personalized medical image segmentation. We evaluated the proposed model on our in-house Nasopharyngeal Carcinoma dataset and the public lung nodule dataset (i.e., LIDC-IDRI). Extensive experiments demonstrated our D-Persona can provide diversified and personalized results at the same time, achieving new SOTA performance for multi-rater medical image segmentation. Our code will be released at https://github.com/ycwu1997/D-Persona.
Authors: Di Wang, Jing Zhang, Minqiang Xu, Lin Liu, Dongsheng Wang, Erzhong Gao, Chengxi Han, Haonan Guo, Bo Du, Dacheng Tao, Liangpei Zhang
Abstract: Foundation models have reshaped the landscape of Remote Sensing (RS) by enhancing various image interpretation tasks. Pretraining is an active research topic, encompassing supervised and self-supervised learning methods to initialize model weights effectively. However, transferring the pretrained models to downstream tasks may encounter task discrepancy due to their formulation of pretraining as image classification or object discrimination tasks. In this study, we explore the Multi-Task Pretraining (MTP) paradigm for RS foundation models to address this issue. Using a shared encoder and task-specific decoder architecture, we conduct multi-task supervised pretraining on the SAMRS dataset, encompassing semantic segmentation, instance segmentation, and rotated object detection. MTP supports both convolutional neural networks and vision transformer foundation models with over 300 million parameters. The pretrained models are finetuned on various RS downstream tasks, such as scene classification, horizontal and rotated object detection, semantic segmentation, and change detection. Extensive experiments across 14 datasets demonstrate the superiority of our models over existing ones of similar size and their competitive performance compared to larger state-of-the-art models, thus validating the effectiveness of MTP.
Authors: Mayura Manawadu, Sieun Park, Soon-Yong Park
Abstract: This study addresses the challenge of accurate 6D pose estimation in Augmented Reality (AR), a critical component for seamlessly integrating virtual objects into real-world environments. Our research primarily addresses the difficulty of estimating 6D poses from uncontrolled RGB images, a common scenario in AR applications, which lacks metadata such as focal length. We propose a novel approach that strategically decomposes the estimation of z-axis translation and focal length, leveraging the neural-render and compare strategy inherent in the FocalPose architecture. This methodology not only streamlines the 6D pose estimation process but also significantly enhances the accuracy of 3D object overlaying in AR settings. Our experimental results demonstrate a marked improvement in 6D pose estimation accuracy, with promising applications in manufacturing and robotics. Here, the precise overlay of AR visualizations and the advancement of robotic vision systems stand to benefit substantially from our findings.
Authors: Chenyang Ma, Kai Lu, Ta-Ying Cheng, Niki Trigoni, Andrew Markham
Abstract: Humans can not only recognize and understand the world in its current state but also envision future scenarios that extend beyond immediate perception. To resemble this profound human capacity, we introduce zero-shot task hallucination -- given a single RGB image of any scene comprising unknown environments and objects, our model can identify potential tasks and imagine their execution in a vivid narrative, realized as a video. We develop a modular pipeline that progressively enhances scene decomposition, comprehension, and reconstruction, incorporating VLM for dynamic interaction and 3D motion planning for object trajectories. Our model can discover diverse tasks, with the generated task videos demonstrating realistic and compelling visual outcomes that are understandable by both machines and humans. Project Page: https://dannymcy.github.io/zeroshot_task_hallucination/
URLs: https://dannymcy.github.io/zeroshot_task_hallucination/
Authors: Xiaoyu Li, Dedong Liu, Lijun Zhao, Yitao Wu, Xian Wu, Jinghan Gao
Abstract: 3D Multi-Object Tracking (MOT) captures stable and comprehensive motion states of surrounding obstacles, essential for robotic perception. However, current 3D trackers face issues with accuracy and latency consistency. In this paper, we propose Fast-Poly, a fast and effective filter-based method for 3D MOT. Building upon our previous work Poly-MOT, Fast-Poly addresses object rotational anisotropy in 3D space, enhances local computation densification, and leverages parallelization technique, improving inference speed and precision. Fast-Poly is extensively tested on two large-scale tracking benchmarks with Python implementation. On the nuScenes dataset, Fast-Poly achieves new state-of-the-art performance with 75.8% AMOTA among all methods and can run at 34.2 FPS on a personal CPU. On the Waymo dataset, Fast-Poly exhibits competitive accuracy with 63.6% MOTA and impressive inference speed (35.5 FPS). The source code is publicly available at https://github.com/lixiaoyu2000/FastPoly.
Authors: Elad Hirsch, Gefen Dawidowicz, Ayellet Tal
Abstract: Generating medical reports for X-ray images presents a significant challenge, particularly in unpaired scenarios where access to paired image-report data for training is unavailable. Previous works have typically learned a joint embedding space for images and reports, necessitating a specific labeling schema for both. We introduce an innovative approach that eliminates the need for consistent labeling schemas, thereby enhancing data accessibility and enabling the use of incompatible datasets. This approach is based on cycle-consistent mapping functions that transform image embeddings into report embeddings, coupled with report auto-encoding for medical report generation. Our model and objectives consider intricate local details and the overarching semantic context within images and reports. This approach facilitates the learning of effective mapping functions, resulting in the generation of coherent reports. It outperforms state-of-the-art results in unpaired chest X-ray report generation, demonstrating improvements in both language and clinical metrics.
Authors: Gowravi Malalur Rajegowda, Yannis Spyridis, Barbara Villarini, Vasileios Argyriou
Abstract: In recent years, there has been an increasing interest in the use of artificial intelligence (AI) and extended reality (XR) in the beauty industry. In this paper, we present an AI-assisted skin care recommendation system integrated into an XR platform. The system uses a convolutional neural network (CNN) to analyse an individual's skin type and recommend personalised skin care products in an immersive and interactive manner. Our methodology involves collecting data from individuals through a questionnaire and conducting skin analysis using a provided facial image in an immersive environment. This data is then used to train the CNN model, which recognises the skin type and existing issues and allows the recommendation engine to suggest personalised skin care products. We evaluate our system in terms of the accuracy of the CNN model, which achieves an average score of 93% in correctly classifying existing skin issues. Being integrated into an XR system, this approach has the potential to significantly enhance the beauty industry by providing immersive and engaging experiences to users, leading to more efficient and consistent skincare routines.
Authors: Zhen Yu, Yang Liu, Qingchao Chen
Abstract: It is essential but challenging to share medical image datasets due to privacy issues, which prohibit building foundation models and knowledge transfer. In this paper, we propose a novel dataset distillation method to condense the original medical image datasets into a synthetic one that preserves useful information for building an analysis model without accessing the original datasets. Existing methods tackle only natural images by randomly matching parts of the training trajectories of the model parameters trained by the whole real datasets. However, through extensive experiments on medical image datasets, the training process is extremely unstable and achieves inferior distillation results. To solve these barriers, we propose to design a novel progressive trajectory matching strategy to improve the training stability for medical image dataset distillation. Additionally, it is observed that improved stability prevents the synthetic dataset diversity and final performance improvements. Therefore, we propose a dynamic overlap mitigation module that improves the synthetic dataset diversity by dynamically eliminating the overlap across different images and retraining parts of the synthetic images for better convergence. Finally, we propose a new medical image dataset distillation benchmark of various modalities and configurations to promote fair evaluations. It is validated that our proposed method achieves 8.33% improvement over previous state-of-the-art methods on average, and 11.7% improvement when ipc=2 (i.e., image per class is 2). Codes and benchmarks will be released.
Authors: Lucas Nunes, Rodrigo Marcuzzi, Benedikt Mersch, Jens Behley, Cyrill Stachniss
Abstract: Computer vision techniques play a central role in the perception stack of autonomous vehicles. Such methods are employed to perceive the vehicle surroundings given sensor data. 3D LiDAR sensors are commonly used to collect sparse 3D point clouds from the scene. However, compared to human perception, such systems struggle to deduce the unseen parts of the scene given those sparse point clouds. In this matter, the scene completion task aims at predicting the gaps in the LiDAR measurements to achieve a more complete scene representation. Given the promising results of recent diffusion models as generative models for images, we propose extending them to achieve scene completion from a single 3D LiDAR scan. Previous works used diffusion models over range images extracted from LiDAR data, directly applying image-based diffusion methods. Distinctly, we propose to directly operate on the points, reformulating the noising and denoising diffusion process such that it can efficiently work at scene scale. Together with our approach, we propose a regularization loss to stabilize the noise predicted during the denoising process. Our experimental evaluation shows that our method can complete the scene given a single LiDAR scan as input, producing a scene with more details compared to state-of-the-art scene completion methods. We believe that our proposed diffusion process formulation can support further research in diffusion models applied to scene-scale point cloud data.
Authors: Davide Alessandro Coccomini, Roberto Caldelli, Claudio Gennaro, Giuseppe Fiameni, Giuseppe Amato, Fabrizio Falchi
Abstract: Deepfake detectors are typically trained on large sets of pristine and generated images, resulting in limited generalization capacity; they excel at identifying deepfakes created through methods encountered during training but struggle with those generated by unknown techniques. This paper introduces a learning approach aimed at significantly enhancing the generalization capabilities of deepfake detectors. Our method takes inspiration from the unique "fingerprints" that image generation processes consistently introduce into the frequency domain. These fingerprints manifest as structured and distinctly recognizable frequency patterns. We propose to train detectors using only pristine images injecting in part of them crafted frequency patterns, simulating the effects of various deepfake generation techniques without being specific to any. These synthetic patterns are based on generic shapes, grids, or auras. We evaluated our approach using diverse architectures across 25 different generation methods. The models trained with our approach were able to perform state-of-the-art deepfake detection, demonstrating also superior generalization capabilities in comparison with previous methods. Indeed, they are untied to any specific generation technique and can effectively identify deepfakes regardless of how they were made.
Authors: Haochen Han, Minnan Luo, Huan Liu, Fang Nan
Abstract: Cross-modal retrieval (CMR) aims to establish interaction between different modalities, among which supervised CMR is emerging due to its flexibility in learning semantic category discrimination. Despite the remarkable performance of previous supervised CMR methods, much of their success can be attributed to the well-annotated data. However, even for unimodal data, precise annotation is expensive and time-consuming, and it becomes more challenging with the multimodal scenario. In practice, massive multimodal data are collected from the Internet with coarse annotation, which inevitably introduces noisy labels. Training with such misleading labels would bring two key challenges -- enforcing the multimodal samples to \emph{align incorrect semantics} and \emph{widen the heterogeneous gap}, resulting in poor retrieval performance. To tackle these challenges, this work proposes UOT-RCL, a Unified framework based on Optimal Transport (OT) for Robust Cross-modal Retrieval. First, we propose a semantic alignment based on partial OT to progressively correct the noisy labels, where a novel cross-modal consistent cost function is designed to blend different modalities and provide precise transport cost. Second, to narrow the discrepancy in multi-modal data, an OT-based relation alignment is proposed to infer the semantic-level cross-modal matching. Both of these two components leverage the inherent correlation among multi-modal data to facilitate effective cost function. The experiments on three widely-used cross-modal retrieval datasets demonstrate that our UOT-RCL surpasses the state-of-the-art approaches and significantly improves the robustness against noisy labels.
Authors: Th\'eophane Vallaeys, Mustafa Shukor, Matthieu Cord, Jakob Verbeek
Abstract: The abilities of large language models (LLMs) have recently progressed to unprecedented levels, paving the way to novel applications in a wide variety of areas. In computer vision, LLMs can be used to prime vision-language tasks such image captioning and visual question answering when coupled with pre-trained vision backbones. While different approaches have been explored to interface LLMs with ``perceptual backbones'' that process, e.g., visual or audio data, they are often explored for different tasks, different datasets, and using different perceptual backbones and language models, hindering direct comparison of the interfacing mechanisms. To remedy this lack of comparability between methods, we present an extensive experimental evaluation of different interfacing mechanisms, across multiple tasks (including image, video, and audio captioning as well as visual question answering), datasets and backbones, paying special attention to low-data settings. We find improved performance using existing mechanisms over state-of-the-art results, and identify a new interfacing mechanism that yields (near) optimal results across different tasks, while obtaining a 4x reduction in training time.
Authors: Yumeng Li, William Beluch, Margret Keuper, Dan Zhang, Anna Khoreva
Abstract: Despite tremendous progress in the field of text-to-video (T2V) synthesis, open-sourced T2V diffusion models struggle to generate longer videos with dynamically varying and evolving content. They tend to synthesize quasi-static videos, ignoring the necessary visual change-over-time implied in the text prompt. At the same time, scaling these models to enable longer, more dynamic video synthesis often remains computationally intractable. To address this challenge, we introduce the concept of Generative Temporal Nursing (GTN), where we aim to alter the generative process on the fly during inference to improve control over the temporal dynamics and enable generation of longer videos. We propose a method for GTN, dubbed VSTAR, which consists of two key ingredients: 1) Video Synopsis Prompting (VSP) - automatic generation of a video synopsis based on the original single prompt leveraging LLMs, which gives accurate textual guidance to different visual states of longer videos, and 2) Temporal Attention Regularization (TAR) - a regularization technique to refine the temporal attention units of the pre-trained T2V diffusion models, which enables control over the video dynamics. We experimentally showcase the superiority of the proposed approach in generating longer, visually appealing videos over existing open-sourced T2V models. We additionally analyze the temporal attention maps realized with and without VSTAR, demonstrating the importance of applying our method to mitigate neglect of the desired visual change over time.
Authors: Jinmin Li, Kuofeng Gao, Yang Bai, Jingyun Zhang, Shu-tao Xia, Yisen Wang
Abstract: Despite the remarkable performance of video-based large language models (LLMs), their adversarial threat remains unexplored. To fill this gap, we propose the first adversarial attack tailored for video-based LLMs by crafting flow-based multi-modal adversarial perturbations on a small fraction of frames within a video, dubbed FMM-Attack. Extensive experiments show that our attack can effectively induce video-based LLMs to generate incorrect answers when videos are added with imperceptible adversarial perturbations. Intriguingly, our FMM-Attack can also induce garbling in the model output, prompting video-based LLMs to hallucinate. Overall, our observations inspire a further understanding of multi-modal robustness and safety-related feature alignment across different modalities, which is of great importance for various large multi-modal models. Our code is available at https://github.com/THU-Kingmin/FMM-Attack.
Authors: Shicai Wei Chunbo Luo Yang Luo
Abstract: Logit knowledge distillation attracts increasing attention due to its practicality in recent studies. However, it often suffers inferior performance compared to the feature knowledge distillation. In this paper, we argue that existing logit-based methods may be sub-optimal since they only leverage the global logit output that couples multiple semantic knowledge. This may transfer ambiguous knowledge to the student and mislead its learning. To this end, we propose a simple but effective method, i.e., Scale Decoupled Distillation (SDD), for logit knowledge distillation. SDD decouples the global logit output into multiple local logit outputs and establishes distillation pipelines for them. This helps the student to mine and inherit fine-grained and unambiguous logit knowledge. Moreover, the decoupled knowledge can be further divided into consistent and complementary logit knowledge that transfers the semantic information and sample ambiguity, respectively. By increasing the weight of complementary parts, SDD can guide the student to focus more on ambiguous samples, improving its discrimination ability. Extensive experiments on several benchmark datasets demonstrate the effectiveness of SDD for wide teacher-student pairs, especially in the fine-grained classification task. Code is available at: https://github.com/shicaiwei123/SDD-CVPR2024
Authors: Junho Kim, Yeon Ju Kim, Yong Man Ro
Abstract: This paper presents a way of enhancing the reliability of Large Multimodal Models (LMMs) in addressing hallucination effects, where models generate incorrect or unrelated responses. Without additional instruction tuning paradigm, we introduce Counterfactual Inception, a novel method that implants counterfactual thoughts into LMMs using carefully chosen, misaligned counterfactual keywords. This method is grounded in the concept of counterfactual thinking, a cognitive process where humans consider alternative realities and outcomes. By applying this human-like reasoning mechanism to LMMs, we aim to reduce hallucination effects and improve the models' trustworthiness. We also propose Dual-modality Verification Process (DVP), a rigorous framework for selecting optimal counterfactual keywords to trigger counterfactual thinking into LMMs, concurrently considering visual and linguistic context. Our extensive experiments across various LMMs, including both open-source and proprietary models, corroborate that our method significantly mitigates hallucination phenomena across different datasets.
Authors: Bowen Zhang, Tianyu Yang, Yu Li, Lei Zhang, Xi Zhao
Abstract: 3D generation has witnessed significant advancements, yet efficiently producing high-quality 3D assets from a single image remains challenging. In this paper, we present a triplane autoencoder, which encodes 3D models into a compact triplane latent space to effectively compress both the 3D geometry and texture information. Within the autoencoder framework, we introduce a 3D-aware cross-attention mechanism, which utilizes low-resolution latent representations to query features from a high-resolution 3D feature volume, thereby enhancing the representation capacity of the latent space. Subsequently, we train a diffusion model on this refined latent space. In contrast to solely relying on image embedding for 3D generation, our proposed method advocates for the simultaneous utilization of both image embedding and shape embedding as conditions. Specifically, the shape embedding is estimated via a diffusion prior model conditioned on the image embedding. Through comprehensive experiments, we demonstrate that our method outperforms state-of-the-art algorithms, achieving superior performance while requiring less training data and time. Our approach enables the generation of high-quality 3D assets in merely 7 seconds on a single A100 GPU.
Authors: Siying Cui, Jiankang Deng, Jia Guo, Xiang An, Yongle Zhao, Xinyu Wei, Ziyong Feng
Abstract: Leveraging Stable Diffusion for the generation of personalized portraits has emerged as a powerful and noteworthy tool, enabling users to create high-fidelity, custom character avatars based on their specific prompts. However, existing personalization methods face challenges, including test-time fine-tuning, the requirement of multiple input images, low preservation of identity, and limited diversity in generated outcomes. To overcome these challenges, we introduce IDAdapter, a tuning-free approach that enhances the diversity and identity preservation in personalized image generation from a single face image. IDAdapter integrates a personalized concept into the generation process through a combination of textual and visual injections and a face identity loss. During the training phase, we incorporate mixed features from multiple reference images of a specific identity to enrich identity-related content details, guiding the model to generate images with more diverse styles, expressions, and angles compared to previous works. Extensive evaluations demonstrate the effectiveness of our method, achieving both diversity and identity fidelity in generated images.
Authors: Konstantinos Alexis, Stella Girtsou, Alexis Apostolakis, Giorgos Giannopoulos, Charalampos Kontoes
Abstract: In this paper we present a deep learning pipeline for next day fire prediction. The next day fire prediction task consists in learning models that receive as input the available information for an area up until a certain day, in order to predict the occurrence of fire for the next day. Starting from our previous problem formulation as a binary classification task on instances (daily snapshots of each area) represented by tabular feature vectors, we reformulate the problem as a semantic segmentation task on images; there, each pixel corresponds to a daily snapshot of an area, while its channels represent the formerly tabular training features. We demonstrate that this problem formulation, built within a thorough pipeline achieves state of the art results.
Authors: Jiwoo Chung, Sangeek Hyun, Sang-Heon Shim, Jae-Pil Heo
Abstract: StyleGAN has shown remarkable performance in unconditional image generation. However, its high computational cost poses a significant challenge for practical applications. Although recent efforts have been made to compress StyleGAN while preserving its performance, existing compressed models still lag behind the original model, particularly in terms of sample diversity. To overcome this, we propose a novel channel pruning method that leverages varying sensitivities of channels to latent vectors, which is a key factor in sample diversity. Specifically, by assessing channel importance based on their sensitivities to latent vector perturbations, our method enhances the diversity of samples in the compressed model. Since our method solely focuses on the channel pruning stage, it has complementary benefits with prior training schemes without additional training cost. Extensive experiments demonstrate that our method significantly enhances sample diversity across various datasets. Moreover, in terms of FID scores, our method not only surpasses state-of-the-art by a large margin but also achieves comparable scores with only half training iterations.
Authors: Hangeol Chang, Jinho Chang, Jong Chul Ye
Abstract: Despite recent advancements in text-to-image diffusion models facilitating various image editing techniques, complex text prompts often lead to an oversight of some requests due to a bottleneck in processing text information. To tackle this challenge, we present Ground-A-Score, a simple yet powerful model-agnostic image editing method by incorporating grounding during score distillation. This approach ensures a precise reflection of intricate prompt requirements in the editing outcomes, taking into account the prior knowledge of the object locations within the image. Moreover, the selective application with a new penalty coefficient and contrastive loss helps to precisely target editing areas while preserving the integrity of the objects in the source image. Both qualitative assessments and quantitative analyses confirm that Ground-A-Score successfully adheres to the intricate details of extended and multifaceted prompts, ensuring high-quality outcomes that respect the original image attributes.
Authors: Djamahl Etchegaray, Zi Huang, Tatsuya Harada, Yadan Luo
Abstract: In this work, we tackle the limitations of current LiDAR-based 3D object detection systems, which are hindered by a restricted class vocabulary and the high costs associated with annotating new object classes. Our exploration of open-vocabulary (OV) learning in urban environments aims to capture novel instances using pre-trained vision-language models (VLMs) with multi-sensor data. We design and benchmark a set of four potential solutions as baselines, categorizing them into either top-down or bottom-up approaches based on their input data strategies. While effective, these methods exhibit certain limitations, such as missing novel objects in 3D box estimation or applying rigorous priors, leading to biases towards objects near the camera or of rectangular geometries. To overcome these limitations, we introduce a universal \textsc{Find n' Propagate} approach for 3D OV tasks, aimed at maximizing the recall of novel objects and propagating this detection capability to more distant areas thereby progressively capturing more. In particular, we utilize a greedy box seeker to search against 3D novel boxes of varying orientations and depth in each generated frustum and ensure the reliability of newly identified boxes by cross alignment and density ranker. Additionally, the inherent bias towards camera-proximal objects is alleviated by the proposed remote simulator, which randomly diversifies pseudo-labeled novel instances in the self-training process, combined with the fusion of base samples in the memory bank. Extensive experiments demonstrate a 53% improvement in novel recall across diverse OV settings, VLMs, and 3D detectors. Notably, we achieve up to a 3.97-fold increase in Average Precision (AP) for novel object classes. The source code is made available in the supplementary material.
Authors: Yu Deng, Duomin Wang, Baoyuan Wang
Abstract: In this paper, we propose a novel learning approach for feed-forward one-shot 4D head avatar synthesis. Different from existing methods that often learn from reconstructing monocular videos guided by 3DMM, we employ pseudo multi-view videos to learn a 4D head synthesizer in a data-driven manner, avoiding reliance on inaccurate 3DMM reconstruction that could be detrimental to the synthesis performance. The key idea is to first learn a 3D head synthesizer using synthetic multi-view images to convert monocular real videos into multi-view ones, and then utilize the pseudo multi-view videos to learn a 4D head synthesizer via cross-view self-reenactment. By leveraging a simple vision transformer backbone with motion-aware cross-attentions, our method exhibits superior performance compared to previous methods in terms of reconstruction fidelity, geometry consistency, and motion control accuracy. We hope our method offers novel insights into integrating 3D priors with 2D supervisions for improved 4D head avatar creation.
Authors: Anh-Kiet Duong, Ho\`ang-\^An L\^e, Minh-Tan Pham
Abstract: In the realm of Federated Learning (FL) applied to remote sensing image classification, this study introduces and assesses several innovative communication strategies. Our exploration includes feature-centric communication, pseudo-weight amalgamation, and a combined method utilizing both weights and features. Experiments conducted on two public scene classification datasets unveil the effectiveness of these strategies, showcasing accelerated convergence, heightened privacy, and reduced network information exchange. This research provides valuable insights into the implications of feature-centric communication in FL, offering potential applications tailored for remote sensing scenarios.
Authors: Yuseung Lee, Minhyuk Sung
Abstract: When an image generation process is guided by both a text prompt and spatial cues, such as a set of bounding boxes, do these elements work in harmony, or does one dominate the other? Our analysis of a pretrained image diffusion model that integrates gated self-attention into the U-Net reveals that spatial grounding often outweighs textual grounding due to the sequential flow from gated self-attention to cross-attention. We demonstrate that such bias can be significantly mitigated without sacrificing accuracy in either grounding by simply rewiring the network architecture, changing from sequential to parallel for gated self-attention and cross-attention. This surprisingly simple yet effective solution does not require any fine-tuning of the network but significantly reduces the trade-off between the two groundings. Our experiments demonstrate significant improvements from the original GLIGEN to the rewired version in the trade-off between textual grounding and spatial grounding.
Authors: Yanyuan Qiao, Zheng Yu, Longteng Guo, Sihan Chen, Zijia Zhao, Mingzhen Sun, Qi Wu, Jing Liu
Abstract: Multimodal large language models (MLLMs) have attracted widespread interest and have rich applications. However, the inherent attention mechanism in its Transformer structure requires quadratic complexity and results in expensive computational overhead. Therefore, in this work, we propose VL-Mamba, a multimodal large language model based on state space models, which have been shown to have great potential for long-sequence modeling with fast inference and linear scaling in sequence length. Specifically, we first replace the transformer-based backbone language model such as LLama or Vicuna with the pre-trained Mamba language model. Then, we empirically explore how to effectively apply the 2D vision selective scan mechanism for multimodal learning and the combinations of different vision encoders and variants of pretrained Mamba language models. The extensive experiments on diverse multimodal benchmarks with competitive performance show the effectiveness of our proposed VL-Mamba and demonstrate the great potential of applying state space models for multimodal learning tasks.
Authors: Renkai Wu, Yinghao Liu, Pengchen Liang, Qing Chang
Abstract: In the field of medical image segmentation, variant models based on Convolutional Neural Networks (CNNs) and Visual Transformers (ViTs) as the base modules have been very widely developed and applied. However, CNNs are often limited in their ability to deal with long sequences of information, while the low sensitivity of ViTs to local feature information and the problem of secondary computational complexity limit their development. Recently, the emergence of state-space models (SSMs), especially 2D-selective-scan (SS2D), has had an impact on the longtime dominance of traditional CNNs and ViTs as the foundational modules of visual neural networks. In this paper, we extend the adaptability of SS2D by proposing a High-order Vision Mamba UNet (H-vmunet) for medical image segmentation. Among them, the proposed High-order 2D-selective-scan (H-SS2D) progressively reduces the introduction of redundant information during SS2D operations through higher-order interactions. In addition, the proposed Local-SS2D module improves the learning ability of local features of SS2D at each order of interaction. We conducted comparison and ablation experiments on three publicly available medical image datasets (ISIC2017, Spleen, and CVC-ClinicDB), and the results all demonstrate the strong competitiveness of H-vmunet in medical image segmentation tasks. The code is available from https://github.com/wurenkai/H-vmunet .
Authors: Junjie Chen, Jiebin Yan, Yuming Fang, Li Niu
Abstract: Category-agnostic pose estimation (CAPE) aims to predict keypoints for arbitrary classes given a few support images annotated with keypoints. Existing methods only rely on the features extracted at support keypoints to predict or refine the keypoints on query image, but a few support feature vectors are local and inadequate for CAPE. Considering that human can quickly perceive potential keypoints of arbitrary objects, we propose a novel framework for CAPE based on such potential keypoints (named as meta-points). Specifically, we maintain learnable embeddings to capture inherent information of various keypoints, which interact with image feature maps to produce meta-points without any support. The produced meta-points could serve as meaningful potential keypoints for CAPE. Due to the inevitable gap between inherency and annotation, we finally utilize the identities and details offered by support keypoints to assign and refine meta-points to desired keypoints in query image. In addition, we propose a progressive deformable point decoder and a slacked regression loss for better prediction and supervision. Our novel framework not only reveals the inherency of keypoints but also outperforms existing methods of CAPE. Comprehensive experiments and in-depth studies on large-scale MP-100 dataset demonstrate the effectiveness of our framework.
Authors: Hiroki Azuma, Yusuke Matsui, Atsuto Maki
Abstract: Deep learning models achieve high accuracy in segmentation tasks among others, yet domain shift often degrades the models' performance, which can be critical in real-world scenarios where no target images are available. This paper proposes a zero-shot domain adaptation method based on diffusion models, called ZoDi, which is two-fold by the design: zero-shot image transfer and model adaptation. First, we utilize an off-the-shelf diffusion model to synthesize target-like images by transferring the domain of source images to the target domain. In this we specifically try to maintain the layout and content by utilising layout-to-image diffusion models with stochastic inversion. Secondly, we train the model using both source images and synthesized images with the original segmentation maps while maximizing the feature similarity of images from the two domains to learn domain-robust representations. Through experiments we show benefits of ZoDi in the task of image segmentation over state-of-the-art methods. It is also more applicable than existing CLIP-based methods because it assumes no specific backbone or models, and it enables to estimate the model's performance without target images by inspecting generated images. Our implementation will be publicly available.
Authors: Florian Strohm, Mihai B\^ace, Andreas Bulling
Abstract: Reusable embeddings of user behaviour have shown significant performance improvements for the personalised saliency prediction task. However, prior works require explicit user characteristics and preferences as input, which are often difficult to obtain. We present a novel method to extract user embeddings from pairs of natural images and corresponding saliency maps generated from a small amount of user-specific eye tracking data. At the core of our method is a Siamese convolutional neural encoder that learns the user embeddings by contrasting the image and personal saliency map pairs of different users. Evaluations on two public saliency datasets show that the generated embeddings have high discriminative power, are effective at refining universal saliency maps to the individual users, and generalise well across users and images. Finally, based on our model's ability to encode individual user characteristics, our work points towards other applications that can benefit from reusable embeddings of gaze behaviour.
Authors: R. Gnana Praveen, Jahangir Alam
Abstract: Multi-modal emotion recognition has recently gained a lot of attention since it can leverage diverse and complementary relationships over multiple modalities, such as audio, visual, and text. Most state-of-the-art methods for multimodal fusion rely on recurrent networks or conventional attention mechanisms that do not effectively leverage the complementary nature of the modalities. In this paper, we focus on dimensional emotion recognition based on the fusion of facial, vocal, and text modalities extracted from videos. Specifically, we propose a recursive cross-modal attention (RCMA) to effectively capture the complementary relationships across the modalities in a recursive fashion. The proposed model is able to effectively capture the inter-modal relationships by computing the cross-attention weights across the individual modalities and the joint representation of the other two modalities. To further improve the inter-modal relationships, the obtained attended features of the individual modalities are again fed as input to the cross-modal attention to refine the feature representations of the individual modalities. In addition to that, we have used Temporal convolution networks (TCNs) to capture the temporal modeling (intra-modal relationships) of the individual modalities. By deploying the TCNs as well cross-modal attention in a recursive fashion, we are able to effectively capture both intra- and inter-modal relationships across the audio, visual, and text modalities. Experimental results on validation-set videos from the AffWild2 dataset indicate that our proposed fusion model is able to achieve significant improvement over the baseline for the sixth challenge of Affective Behavior Analysis in-the-Wild 2024 (ABAW6) competition.
Authors: Jianhao Xie, Ruofan Liao, Ziang Zhang, Sida Yi, Yuesheng Zhu, Guibo Luo
Abstract: Detecting polyps through colonoscopy is an important task in medical image segmentation, which provides significant assistance and reference value for clinical surgery. However, accurate segmentation of polyps is a challenging task due to two main reasons. Firstly, polyps exhibit various shapes and colors. Secondly, the boundaries between polyps and their normal surroundings are often unclear. Additionally, significant differences between different datasets lead to limited generalization capabilities of existing methods. To address these issues, we propose a segmentation model based on Prompt-Mamba, which incorporates the latest Vision-Mamba and prompt technologies. Compared to previous models trained on the same dataset, our model not only maintains high segmentation accuracy on the validation part of the same dataset but also demonstrates superior accuracy on unseen datasets, exhibiting excellent generalization capabilities. Notably, we are the first to apply the Vision-Mamba architecture to polyp segmentation and the first to utilize prompt technology in a polyp segmentation model. Our model efficiently accomplishes segmentation tasks, surpassing previous state-of-the-art methods by an average of 5% across six datasets. Furthermore, we have developed multiple versions of our model with scaled parameter counts, achieving better performance than previous models even with fewer parameters. Our code and trained weights will be released soon.
Authors: Shijie Zhang, Boyan Jiang, Keke He, Junwei Zhu, Ying Tai, Chengjie Wang, Yinda Zhang, Yanwei Fu
Abstract: Pixel2Mesh (P2M) is a classical approach for reconstructing 3D shapes from a single color image through coarse-to-fine mesh deformation. Although P2M is capable of generating plausible global shapes, its Graph Convolution Network (GCN) often produces overly smooth results, causing the loss of fine-grained geometry details. Moreover, P2M generates non-credible features for occluded regions and struggles with the domain gap from synthetic data to real-world images, which is a common challenge for single-view 3D reconstruction methods. To address these challenges, we propose a novel Transformer-boosted architecture, named T-Pixel2Mesh, inspired by the coarse-to-fine approach of P2M. Specifically, we use a global Transformer to control the holistic shape and a local Transformer to progressively refine the local geometry details with graph-based point upsampling. To enhance real-world reconstruction, we present the simple yet effective Linear Scale Search (LSS), which serves as prompt tuning during the input preprocessing. Our experiments on ShapeNet demonstrate state-of-the-art performance, while results on real-world data show the generalization capability.
Authors: Zixuan Wang, Jia Jia, Shikun Sun, Haozhe Wu, Rong Han, Zhenyu Li, Di Tang, Jiaqing Zhou, Jiebo Luo
Abstract: Choreographers determine what the dances look like, while cameramen determine the final presentation of dances. Recently, various methods and datasets have showcased the feasibility of dance synthesis. However, camera movement synthesis with music and dance remains an unsolved challenging problem due to the scarcity of paired data. Thus, we present DCM, a new multi-modal 3D dataset, which for the first time combines camera movement with dance motion and music audio. This dataset encompasses 108 dance sequences (3.2 hours) of paired dance-camera-music data from the anime community, covering 4 music genres. With this dataset, we uncover that dance camera movement is multifaceted and human-centric, and possesses multiple influencing factors, making dance camera synthesis a more challenging task compared to camera or dance synthesis alone. To overcome these difficulties, we propose DanceCamera3D, a transformer-based diffusion model that incorporates a novel body attention loss and a condition separation strategy. For evaluation, we devise new metrics measuring camera movement quality, diversity, and dancer fidelity. Utilizing these metrics, we conduct extensive experiments on our DCM dataset, providing both quantitative and qualitative evidence showcasing the effectiveness of our DanceCamera3D model. Code and video demos are available at https://github.com/Carmenw1203/DanceCamera3D-Official.
URLs: https://github.com/Carmenw1203/DanceCamera3D-Official.
Authors: Yuyang Shu, Michael E. Bain
Abstract: Humans see low and high spatial frequency components at the same time, and combine the information from both to form a visual scene. Drawing on this neuroscientific inspiration, we propose an altered Vision Transformer architecture where patches from scaled down versions of the input image are added to the input of the first Transformer Encoder layer. We name this model Retina Vision Transformer (RetinaViT) due to its inspiration from the human visual system. Our experiments show that when trained on the ImageNet-1K dataset with a moderate configuration, RetinaViT achieves a 3.3% performance improvement over the original ViT. We hypothesize that this improvement can be attributed to the inclusion of low spatial frequency components in the input, which improves the ability to capture structural features, and to select and forward important features to deeper layers. RetinaViT thereby opens doors to further investigations into vertical pathways and attention patterns.
Authors: Jun Yu, Zerui Zhang, Zhihong Wei, Gongpeng Zhao, Zhongpeng Cai, Yongqi Wang, Guochen Xie, Jichao Zhu, Wangyuan Zhu
Abstract: Leveraging the synergy of both audio data and visual data is essential for understanding human emotions and behaviors, especially in in-the-wild setting. Traditional methods for integrating such multimodal information often stumble, leading to less-than-ideal outcomes in the task of facial action unit detection. To overcome these shortcomings, we propose a novel approach utilizing audio-visual multimodal data. This method enhances audio feature extraction by leveraging Mel Frequency Cepstral Coefficients (MFCC) and Log-Mel spectrogram features alongside a pre-trained VGGish network. Moreover, this paper adaptively captures fusion features across modalities by modeling the temporal relationships, and ultilizes a pre-trained GPT-2 model for sophisticated context-aware fusion of multimodal information. Our method notably improves the accuracy of AU detection by understanding the temporal and contextual nuances of the data, showcasing significant advancements in the comprehension of intricate scenarios. These findings underscore the potential of integrating temporal dynamics and contextual interpretation, paving the way for future research endeavors.
Authors: Chen Zhao, Tong Zhang, Zheng Dang, Mathieu Salzmann
Abstract: Determining the relative pose of an object between two images is pivotal to the success of generalizable object pose estimation. Existing approaches typically approximate the continuous pose representation with a large number of discrete pose hypotheses, which incurs a computationally expensive process of scoring each hypothesis at test time. By contrast, we present a Deep Voxel Matching Network (DVMNet) that eliminates the need for pose hypotheses and computes the relative object pose in a single pass. To this end, we map the two input RGB images, reference and query, to their respective voxelized 3D representations. We then pass the resulting voxels through a pose estimation module, where the voxels are aligned and the pose is computed in an end-to-end fashion by solving a least-squares problem. To enhance robustness, we introduce a weighted closest voxel algorithm capable of mitigating the impact of noisy voxels. We conduct extensive experiments on the CO3D, LINEMOD, and Objaverse datasets, demonstrating that our method delivers more accurate relative pose estimates for novel objects at a lower computational cost compared to state-of-the-art methods. Our code is released at: https://github.com/sailor-z/DVMNet/.
Authors: Hongjun Wang, Sagar Vaze, Kai Han
Abstract: Generalized Category Discovery (GCD) aims to classify unlabelled images from both `seen' and `unseen' classes by transferring knowledge from a set of labelled `seen' class images. A key theme in existing GCD approaches is adapting large-scale pre-trained models for the GCD task. An alternate perspective, however, is to adapt the data representation itself for better alignment with the pre-trained model. As such, in this paper, we introduce a two-stage adaptation approach termed SPTNet, which iteratively optimizes model parameters (i.e., model-finetuning) and data parameters (i.e., prompt learning). Furthermore, we propose a novel spatial prompt tuning method (SPT) which considers the spatial property of image data, enabling the method to better focus on object parts, which can transfer between seen and unseen classes. We thoroughly evaluate our SPTNet on standard benchmarks and demonstrate that our method outperforms existing GCD methods. Notably, we find our method achieves an average accuracy of 61.4% on the SSB, surpassing prior state-of-the-art methods by approximately 10%. The improvement is particularly remarkable as our method yields extra parameters amounting to only 0.117% of those in the backbone architecture. Project page: https://visual-ai.github.io/sptnet.
Authors: Tran-Vu La, Minh-Tan Pham, Marco Chini
Abstract: Ship detection from satellite imagery using Deep Learning (DL) is an indispensable solution for maritime surveillance. However, applying DL models trained on one dataset to others having differences in spatial resolution and radiometric features requires many adjustments. To overcome this issue, this paper focused on the DL models trained on datasets that consist of different optical images and a combination of radar and optical data. When dealing with a limited number of training images, the performance of DL models via this approach was satisfactory. They could improve 5-20% of average precision, depending on the optical images tested. Likewise, DL models trained on the combined optical and radar dataset could be applied to both optical and radar images. Our experiments showed that the models trained on an optical dataset could be used for radar images, while those trained on a radar dataset offered very poor scores when applied to optical images.
Authors: Danqing Ma, Shaojie Li, Bo Dang, Hengyi Zang, Xinqi Dong
Abstract: Transmission line detection technology is crucial for automatic monitoring and ensuring the safety of electrical facilities. The YOLOv5 series is currently one of the most advanced and widely used methods for object detection. However, it faces inherent challenges, such as high computational load on devices and insufficient detection accuracy. To address these concerns, this paper presents an enhanced lightweight YOLOv5 technique customized for mobile devices, specifically intended for identifying objects associated with transmission lines. The C3Ghost module is integrated into the convolutional network of YOLOv5 to reduce floating point operations per second (FLOPs) in the feature channel fusion process and improve feature expression performance. In addition, a FasterNet module is introduced to replace the c3 module in the YOLOv5 Backbone. The FasterNet module uses Partial Convolutions to process only a portion of the input channels, improving feature extraction efficiency and reducing computational overhead. To address the imbalance between simple and challenging samples in the dataset and the diversity of aspect ratios of bounding boxes, the wIoU v3 LOSS is adopted as the loss function. To validate the performance of the proposed approach, Experiments are conducted on a custom dataset of transmission line poles. The results show that the proposed model achieves a 1% increase in detection accuracy, a 13% reduction in FLOPs, and a 26% decrease in model parameters compared to the existing YOLOv5.In the ablation experiment, it was also discovered that while the Fastnet module and the CSghost module improved the precision of the original YOLOv5 baseline model, they caused a decrease in the mAP@.5-.95 metric. However, the improvement of the wIoUv3 loss function significantly mitigated the decline of the mAP@.5-.95 metric.
Authors: Seongjae Min, Junseok Yang, Sangjun Lim, Junyong Lee, Sangwon Lee, Sejoon Lim
Abstract: In recent years, deep learning has achieved innovative advancements in various fields, including the analysis of human emotions and behaviors. Initiatives such as the Affective Behavior Analysis in-the-wild (ABAW) competition have been particularly instrumental in driving research in this area by providing diverse and challenging datasets that enable precise evaluation of complex emotional states. This study leverages the Vision Transformer (ViT) and Transformer models to focus on the estimation of Valence-Arousal (VA), which signifies the positivity and intensity of emotions, recognition of various facial expressions, and detection of Action Units (AU) representing fundamental muscle movements. This approach transcends traditional Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) based methods, proposing a new Transformer-based framework that maximizes the understanding of temporal and spatial features. The core contributions of this research include the introduction of a learning technique through random frame masking and the application of Focal loss adapted for imbalanced data, enhancing the accuracy and applicability of emotion and behavior analysis in real-world settings. This approach is expected to contribute to the advancement of emotional computing and deep learning methodologies.
Authors: Fu-Yun Wang, Xiaoshi Wu, Zhaoyang Huang, Xiaoyu Shi, Dazhong Shen, Guanglu Song, Yu Liu, Hongsheng Li
Abstract: Video outpainting is a challenging task, aiming at generating video content outside the viewport of the input video while maintaining inter-frame and intra-frame consistency. Existing methods fall short in either generation quality or flexibility. We introduce MOTIA Mastering Video Outpainting Through Input-Specific Adaptation, a diffusion-based pipeline that leverages both the intrinsic data-specific patterns of the source video and the image/video generative prior for effective outpainting. MOTIA comprises two main phases: input-specific adaptation and pattern-aware outpainting. The input-specific adaptation phase involves conducting efficient and effective pseudo outpainting learning on the single-shot source video. This process encourages the model to identify and learn patterns within the source video, as well as bridging the gap between standard generative processes and outpainting. The subsequent phase, pattern-aware outpainting, is dedicated to the generalization of these learned patterns to generate outpainting outcomes. Additional strategies including spatial-aware insertion and noise travel are proposed to better leverage the diffusion model's generative prior and the acquired video patterns from source videos. Extensive evaluations underscore MOTIA's superiority, outperforming existing state-of-the-art methods in widely recognized benchmarks. Notably, these advancements are achieved without necessitating extensive, task-specific tuning.
Authors: Aymene Berriche, Mehdi Adjal Zakaria, Riyadh Baghdadi
Abstract: Deep hashing techniques have emerged as the predominant approach for efficient image retrieval. Traditionally, these methods utilize pre-trained convolutional neural networks (CNNs) such as AlexNet and VGG-16 as feature extractors. However, the increasing complexity of datasets poses challenges for these backbone architectures in capturing meaningful features essential for effective image retrieval. In this study, we explore the efficacy of employing high-resolution features learned through state-of-the-art techniques for image retrieval tasks. Specifically, we propose a novel methodology that utilizes High-Resolution Networks (HRNets) as the backbone for the deep hashing task, termed High-Resolution Hashing Network (HHNet). Our approach demonstrates superior performance compared to existing methods across all tested benchmark datasets, including CIFAR-10, NUS-WIDE, MS COCO, and ImageNet. This performance improvement is more pronounced for complex datasets, which highlights the need to learn high-resolution features for intricate image retrieval tasks. Furthermore, we conduct a comprehensive analysis of different HRNet configurations and provide insights into the optimal architecture for the deep hashing task
Authors: Diwei Wang, Kun Yuan, Candice Muller, Fr\'ed\'eric Blanc, Nicolas Padoy, Hyewon Seo
Abstract: We present a knowledge augmentation strategy for assessing the diagnostic groups and gait impairment from monocular gait videos. Based on a large-scale pre-trained Vision Language Model (VLM), our model learns and improves visual, textual, and numerical representations of patient gait videos, through a collective learning across three distinct modalities: gait videos, class-specific descriptions, and numerical gait parameters. Our specific contributions are two-fold: First, we adopt a knowledge-aware prompt tuning strategy to utilize the class-specific medical description in guiding the text prompt learning. Second, we integrate the paired gait parameters in the form of numerical texts to enhance the numeracy of the textual representation. Results demonstrate that our model not only significantly outperforms state-of-the-art (SOTA) in video-based classification tasks but also adeptly decodes the learned class-specific text features into natural language descriptions using the vocabulary of quantitative gait parameters. The code and the model will be made available at our project page.
Authors: Yuyi Zhang, Yuanzhi Zhu, Dezhi Peng, Peirong Zhang, Zhenhua Yang, Zhibo Yang, Cong Yao, Lianwen Jin
Abstract: Text recognition, especially for complex scripts like Chinese, faces unique challenges due to its intricate character structures and vast vocabulary. Traditional one-hot encoding methods struggle with the representation of hierarchical radicals, recognition of Out-Of-Vocabulary (OOV) characters, and on-device deployment due to their computational intensity. To address these challenges, we propose HierCode, a novel and lightweight codebook that exploits the innate hierarchical nature of Chinese characters. HierCode employs a multi-hot encoding strategy, leveraging hierarchical binary tree encoding and prototype learning to create distinctive, informative representations for each character. This approach not only facilitates zero-shot recognition of OOV characters by utilizing shared radicals and structures but also excels in line-level recognition tasks by computing similarity with visual features, a notable advantage over existing methods. Extensive experiments across diverse benchmarks, including handwritten, scene, document, web, and ancient text, have showcased HierCode's superiority for both conventional and zero-shot Chinese character or text recognition, exhibiting state-of-the-art performance with significantly fewer parameters and fast inference speed.
Authors: Giulia Rizzoli, Matteo Caligiuri, Donald Shenaj, Francesco Barbato, Pietro Zanuttigh
Abstract: In Federated Learning (FL), multiple clients collaboratively train a global model without sharing private data. In semantic segmentation, the Federated source Free Domain Adaptation (FFreeDA) setting is of particular interest, where clients undergo unsupervised training after supervised pretraining at the server side. While few recent works address FL for autonomous vehicles, intrinsic real-world challenges such as the presence of adverse weather conditions and the existence of different autonomous agents are still unexplored. To bridge this gap, we address both problems and introduce a new federated semantic segmentation setting where both car and drone clients co-exist and collaborate. Specifically, we propose a novel approach for this setting which exploits a batch-norm weather-aware strategy to dynamically adapt the model to the different weather conditions, while hyperbolic space prototypes are used to align the heterogeneous client representations. Finally, we introduce FLYAWARE, the first semantic segmentation dataset with adverse weather data for aerial vehicles.
Authors: Ji\v{r}\'i Mayer, Milan Straka, Jan Haji\v{c} jr., Pavel Pecina
Abstract: The majority of recent progress in Optical Music Recognition (OMR) has been achieved with Deep Learning methods, especially models following the end-to-end paradigm, reading input images and producing a linear sequence of tokens. Unfortunately, many music scores, especially piano music, cannot be easily converted to a linear sequence. This has led OMR researchers to use custom linearized encodings, instead of broadly accepted structured formats for music notation. Their diversity makes it difficult to compare the performance of OMR systems directly. To bring recent OMR model progress closer to useful results: (a) We define a sequential format called Linearized MusicXML, allowing to train an end-to-end model directly and maintaining close cohesion and compatibility with the industry-standard MusicXML format. (b) We create a dev and test set for benchmarking typeset OMR with MusicXML ground truth based on the OpenScore Lieder corpus. They contain 1,438 and 1,493 pianoform systems, each with an image from IMSLP. (c) We train and fine-tune an end-to-end model to serve as a baseline on the dataset and employ the TEDn metric to evaluate the model. We also test our model against the recently published synthetic pianoform dataset GrandStaff and surpass the state-of-the-art results.
Authors: Nicholas Bai, Rahul A. Iyer, Tuomas Oikarinen, Tsui-Wei Weng
Abstract: In this paper, we propose Describe-and-Dissect (DnD), a novel method to describe the roles of hidden neurons in vision networks. DnD utilizes recent advancements in multimodal deep learning to produce complex natural language descriptions, without the need for labeled training data or a predefined set of concepts to choose from. Additionally, DnD is training-free, meaning we don't train any new models and can easily leverage more capable general purpose models in the future. We have conducted extensive qualitative and quantitative analysis to show that DnD outperforms prior work by providing higher quality neuron descriptions. Specifically, our method on average provides the highest quality labels and is more than 2 times as likely to be selected as the best explanation for a neuron than the best baseline.
Authors: Mohammadhossein Bahari, Saeed Saadatnejad, Amirhossein Asgari Farsangi, Seyed-Mohsen Moosavi-Dezfooli, Alexandre Alahi
Abstract: Trajectory prediction plays an essential role in autonomous vehicles. While numerous strategies have been developed to enhance the robustness of trajectory prediction models, these methods are predominantly heuristic and do not offer guaranteed robustness against adversarial attacks and noisy observations. In this work, we propose a certification approach tailored for the task of trajectory prediction. To this end, we address the inherent challenges associated with trajectory prediction, including unbounded outputs, and mutli-modality, resulting in a model that provides guaranteed robustness. Furthermore, we integrate a denoiser into our method to further improve the performance. Through comprehensive evaluations, we demonstrate the effectiveness of the proposed technique across various baselines and using standard trajectory prediction datasets. The code will be made available online: https://s-attack.github.io/
Authors: Ming Gui, Johannes S. Fischer, Ulrich Prestel, Pingchuan Ma, Dmytro Kotovenko, Olga Grebenkova, Stefan Andreas Baumann, Vincent Tao Hu, Bj\"orn Ommer
Abstract: Monocular depth estimation is crucial for numerous downstream vision tasks and applications. Current discriminative approaches to this problem are limited due to blurry artifacts, while state-of-the-art generative methods suffer from slow sampling due to their SDE nature. Rather than starting from noise, we seek a direct mapping from input image to depth map. We observe that this can be effectively framed using flow matching, since its straight trajectories through solution space offer efficiency and high quality. Our study demonstrates that a pre-trained image diffusion model can serve as an adequate prior for a flow matching depth model, allowing efficient training on only synthetic data to generalize to real images. We find that an auxiliary surface normals loss further improves the depth estimates. Due to the generative nature of our approach, our model reliably predicts the confidence of its depth estimates. On standard benchmarks of complex natural scenes, our lightweight approach exhibits state-of-the-art performance at favorable low computational cost despite only being trained on little synthetic data.
Authors: Lauren Okamoto, Paritosh Parmar
Abstract: Action quality assessment (AQA) applies computer vision to quantitatively assess the performance or execution of a human action. Current AQA approaches are end-to-end neural models, which lack transparency and tend to be biased because they are trained on subjective human judgements as ground-truth. To address these issues, we introduce a neuro-symbolic paradigm for AQA, which uses neural networks to abstract interpretable symbols from video data and makes quality assessments by applying rules to those symbols. We take diving as the case study. We found that domain experts prefer our system and find it more informative than purely neural approaches to AQA in diving. Our system also achieves state-of-the-art action recognition and temporal segmentation, and automatically generates a detailed report that breaks the dive down into its elements and provides objective scoring with visual evidence. As verified by a group of domain experts, this report may be used to assist judges in scoring, help train judges, and provide feedback to divers. We will open-source all of our annotated training data and code for ease of reproducibility.
Authors: Jingxi Chen, Brandon Y. Feng, Haoming Cai, Mingyang Xie, Christopher Metzler, Cornelia Fermuller, Yiannis Aloimonos
Abstract: This paper addresses the novel challenge of ``rewinding'' time from a single captured image to recover the fleeting moments missed just before the shutter button is pressed. This problem poses a significant challenge in computer vision and computational photography, as it requires predicting plausible pre-capture motion from a single static frame, an inherently ill-posed task due to the high degree of freedom in potential pixel movements. We overcome this challenge by leveraging the emerging technology of neuromorphic event cameras, which capture motion information with high temporal resolution, and integrating this data with advanced image-to-video diffusion models. Our proposed framework introduces an event motion adaptor conditioned on event camera data, guiding the diffusion model to generate videos that are visually coherent and physically grounded in the captured events. Through extensive experimentation, we demonstrate the capability of our approach to synthesize high-quality videos that effectively ``rewind'' time, showcasing the potential of combining event camera technology with generative models. Our work opens new avenues for research at the intersection of computer vision, computational photography, and generative modeling, offering a forward-thinking solution to capturing missed moments and enhancing future consumer cameras and smartphones. Please see the project page at https://timerewind.github.io/ for video results and code release.
Authors: Vincent Tao Hu, Stefan Andreas Baumann, Ming Gui, Olga Grebenkova, Pingchuan Ma, Johannes Fischer, Bjorn Ommer
Abstract: The diffusion model has long been plagued by scalability and quadratic complexity issues, especially within transformer-based structures. In this study, we aim to leverage the long sequence modeling capability of a State-Space Model called Mamba to extend its applicability to visual data generation. Firstly, we identify a critical oversight in most current Mamba-based vision methods, namely the lack of consideration for spatial continuity in the scan scheme of Mamba. Secondly, building upon this insight, we introduce a simple, plug-and-play, zero-parameter method named Zigzag Mamba, which outperforms Mamba-based baselines and demonstrates improved speed and memory utilization compared to transformer-based baselines. Lastly, we integrate Zigzag Mamba with the Stochastic Interpolant framework to investigate the scalability of the model on large-resolution visual datasets, such as FacesHQ $1024\times 1024$ and UCF101, MultiModal-CelebA-HQ, and MS COCO $256\times 256$. Code will be released at https://taohu.me/zigma/
URLs: https://taohu.me/zigma/
Authors: Yang Yang, Wenhai Wang, Zhe Chen, Jifeng Dai, Liang Zheng
Abstract: Bounding boxes uniquely characterize object detection, where a good detector gives accurate bounding boxes of categories of interest. However, in the real-world where test ground truths are not provided, it is non-trivial to find out whether bounding boxes are accurate, thus preventing us from assessing the detector generalization ability. In this work, we find under feature map dropout, good detectors tend to output bounding boxes whose locations do not change much, while bounding boxes of poor detectors will undergo noticeable position changes. We compute the box stability score (BoS score) to reflect this stability. Specifically, given an image, we compute a normal set of bounding boxes and a second set after feature map dropout. To obtain BoS score, we use bipartite matching to find the corresponding boxes between the two sets and compute the average Intersection over Union (IoU) across the entire test set. We contribute to finding that BoS score has a strong, positive correlation with detection accuracy measured by mean average precision (mAP) under various test environments. This relationship allows us to predict the accuracy of detectors on various real-world test sets without accessing test ground truths, verified on canonical detection tasks such as vehicle detection and pedestrian detection. Code and data are available at https://github.com/YangYangGirl/BoS.
Authors: Ruozhen He, Paola Cascante-Bonilla, Ziyan Yang, Alexander C. Berg, Vicente Ordonez
Abstract: We introduce SynGround, a novel framework that combines data-driven learning and knowledge transfer from various large-scale pretrained models to enhance the visual grounding capabilities of a pretrained vision-and-language model. The knowledge transfer from the models initiates the generation of image descriptions through an image description generator. These descriptions serve dual purposes: they act as prompts for synthesizing images through a text-to-image generator, and as queries for synthesizing text, from which phrases are extracted using a large language model. Finally, we leverage an open-vocabulary object detector to generate synthetic bounding boxes for the synthetic images and texts. We finetune a pretrained vision-and-language model on this dataset by optimizing a mask-attention consistency objective that aligns region annotations with gradient-based model explanations. The resulting model improves the grounding capabilities of an off-the-shelf vision-and-language model. Particularly, SynGround improves the pointing game accuracy of ALBEF on the Flickr30k dataset from 79.38% to 87.26%, and on RefCOCO+ Test A from 69.35% to 79.06% and on RefCOCO+ Test B from 53.77% to 63.67%.
Authors: Ziyu Liu, Zeyi Sun, Yuhang Zang, Wei Li, Pan Zhang, Xiaoyi Dong, Yuanjun Xiong, Dahua Lin, Jiaqi Wang
Abstract: CLIP (Contrastive Language-Image Pre-training) uses contrastive learning from noise image-text pairs to excel at recognizing a wide array of candidates, yet its focus on broad associations hinders the precision in distinguishing subtle differences among fine-grained items. Conversely, Multimodal Large Language Models (MLLMs) excel at classifying fine-grained categories, thanks to their substantial knowledge from pre-training on web-level corpora. However, the performance of MLLMs declines with an increase in category numbers, primarily due to growing complexity and constraints of limited context window size. To synergize the strengths of both approaches and enhance the few-shot/zero-shot recognition abilities for datasets characterized by extensive and fine-grained vocabularies, this paper introduces RAR, a Retrieving And Ranking augmented method for MLLMs. We initially establish a multi-modal retriever based on CLIP to create and store explicit memory for different categories beyond the immediate context window. During inference, RAR retrieves the top-k similar results from the memory and uses MLLMs to rank and make the final predictions. Our proposed approach not only addresses the inherent limitations in fine-grained recognition but also preserves the model's comprehensive knowledge base, significantly boosting accuracy across a range of vision-language recognition tasks. Notably, our approach demonstrates a significant improvement in performance on 5 fine-grained visual recognition benchmarks, 11 few-shot image recognition datasets, and the 2 object detection datasets under the zero-shot recognition setting.
Authors: Michael Niemeyer, Fabian Manhardt, Marie-Julie Rakotosaona, Michael Oechsle, Daniel Duckworth, Rama Gosula, Keisuke Tateno, John Bates, Dominik Kaeser, Federico Tombari
Abstract: Recent advances in view synthesis and real-time rendering have achieved photorealistic quality at impressive rendering speeds. While Radiance Field-based methods achieve state-of-the-art quality in challenging scenarios such as in-the-wild captures and large-scale scenes, they often suffer from excessively high compute requirements linked to volumetric rendering. Gaussian Splatting-based methods, on the other hand, rely on rasterization and naturally achieve real-time rendering but suffer from brittle optimization heuristics that underperform on more challenging scenes. In this work, we present RadSplat, a lightweight method for robust real-time rendering of complex scenes. Our main contributions are threefold. First, we use radiance fields as a prior and supervision signal for optimizing point-based scene representations, leading to improved quality and more robust optimization. Next, we develop a novel pruning technique reducing the overall point count while maintaining high quality, leading to smaller and more compact scene representations with faster inference speeds. Finally, we propose a novel test-time filtering approach that further accelerates rendering and allows to scale to larger, house-sized scenes. We find that our method enables state-of-the-art synthesis of complex captures at 900+ FPS.
Authors: Tianwei Xiong, Yue Wu, Enze Xie, Yue Wu, Zhenguo Li, Xihui Liu
Abstract: Text-to-image diffusion models suffer from the risk of generating outdated, copyrighted, incorrect, and biased content. While previous methods have mitigated the issues on a small scale, it is essential to handle them simultaneously in larger-scale real-world scenarios. We propose a two-stage method, Editing Massive Concepts In Diffusion Models (EMCID). The first stage performs memory optimization for each individual concept with dual self-distillation from text alignment loss and diffusion noise prediction loss. The second stage conducts massive concept editing with multi-layer, closed form model editing. We further propose a comprehensive benchmark, named ImageNet Concept Editing Benchmark (ICEB), for evaluating massive concept editing for T2I models with two subtasks, free-form prompts, massive concept categories, and extensive evaluation metrics. Extensive experiments conducted on our proposed benchmark and previous benchmarks demonstrate the superior scalability of EMCID for editing up to 1,000 concepts, providing a practical approach for fast adjustment and re-deployment of T2I diffusion models in real-world applications.
Authors: Hasan Abed Al Kader Hammoud, Tuhin Das, Fabio Pizzati, Philip Torr, Adel Bibi, Bernard Ghanem
Abstract: We explore the impact of training with more diverse datasets, characterized by the number of unique samples, on the performance of self-supervised learning (SSL) under a fixed computational budget. Our findings consistently demonstrate that increasing pretraining data diversity enhances SSL performance, albeit only when the distribution distance to the downstream data is minimal. Notably, even with an exceptionally large pretraining data diversity achieved through methods like web crawling or diffusion-generated data, among other ways, the distribution shift remains a challenge. Our experiments are comprehensive with seven SSL methods using large-scale datasets such as ImageNet and YFCC100M amounting to over 200 GPU days. Code and trained models will be available at https://github.com/hammoudhasan/DiversitySSL .
Authors: Junyao Gao, Xinyang Jiang, Huishuai Zhang, Yifan Yang, Shuguang Dou, Dongsheng Li, Duoqian Miao, Cheng Deng, Cairong Zhao
Abstract: While person Re-identification (Re-ID) has progressed rapidly due to its wide real-world applications, it also causes severe risks of leaking personal information from training data. Thus, this paper focuses on quantifying this risk by membership inference (MI) attack. Most of the existing MI attack algorithms focus on classification models, while Re-ID follows a totally different training and inference paradigm. Re-ID is a fine-grained recognition task with complex feature embedding, and model outputs commonly used by existing MI like logits and losses are not accessible during inference. Since Re-ID focuses on modelling the relative relationship between image pairs instead of individual semantics, we conduct a formal and empirical analysis which validates that the distribution shift of the inter-sample similarity between training and test set is a critical criterion for Re-ID membership inference. As a result, we propose a novel membership inference attack method based on the inter-sample similarity distribution. Specifically, a set of anchor images are sampled to represent the similarity distribution conditioned on a target image, and a neural network with a novel anchor selection module is proposed to predict the membership of the target image. Our experiments validate the effectiveness of the proposed approach on both the Re-ID task and conventional classification task.
Authors: Ruiqing Sun, Delong Yang, Shaohui Zhang, Qun Hao
Abstract: Relying on either deep models or physical models are two mainstream approaches for solving inverse sample reconstruction problems in programmable illumination computational microscopy. Solutions based on physical models possess strong generalization capabilities while struggling with global optimization of inverse problems due to a lack of insufficient physical constraints. In contrast, deep learning methods have strong problem-solving abilities, but their generalization ability is often questioned because of the unclear physical principles. Besides, conventional deep models are difficult to apply to some specific scenes because of the difficulty in acquiring high-quality training data and their limited capacity to generalize across different scenarios. In this paper, to combine the advantages of deep models and physical models together, we propose a hybrid framework consisting of three sub-neural networks (two deep learning networks and one physics-based network). We first obtain a result with rich semantic information through a light deep learning neural network and then use it as the initial value of the physical network to make its output comply with physical process constraints. These two results are then used as the input of a fusion deep learning neural work which utilizes the paired features between the reconstruction results of two different models to further enhance imaging quality. The final result integrates the advantages of both deep models and physical models and can quickly solve the computational reconstruction inverse problem in programmable illumination computational microscopy and achieve better results. We verified the feasibility and effectiveness of the proposed hybrid framework with theoretical analysis and actual experiments on resolution targets and biological samples.
Authors: Ahmed F. AbouElhamayed, Susanne Balle, Deshanand Singh, Mohamed S. Abdelfattah
Abstract: Deep neural network (DNN) inference has become an important part of many data-center workloads. This has prompted focused efforts to design ever-faster deep learning accelerators such as GPUs and TPUs. However, an end-to-end DNN-based vision application contains more than just DNN inference, including input decompression, resizing, sampling, normalization, and data transfer. In this paper, we perform a thorough evaluation of computer vision inference requests performed on a throughput-optimized serving system. We quantify the performance impact of server overheads such as data movement, preprocessing, and message brokers between two DNNs producing outputs at different rates. Our empirical analysis encompasses many computer vision tasks including image classification, segmentation, detection, depth-estimation, and more complex processing pipelines with multiple DNNs. Our results consistently demonstrate that end-to-end application performance can easily be dominated by data processing and data movement functions (up to 56% of end-to-end latency in a medium-sized image, and $\sim$ 80% impact on system throughput in a large image), even though these functions have been conventionally overlooked in deep learning system design. Our work identifies important performance bottlenecks in different application scenarios, achieves 2.25$\times$ better throughput compared to prior work, and paves the way for more holistic deep learning system design.
Authors: Jixiang Luo, Yan Wang, Hongwei Qin
Abstract: Learned Image Compression (LIC) has achieved dramatic progress regarding objective and subjective metrics. MSE-based models aim to improve objective metrics while generative models are leveraged to improve visual quality measured by subjective metrics. However, they all suffer from blurring or deformation at low bit rates, especially at below $0.2bpp$. Besides, deformation on human faces and text is unacceptable for visual quality assessment, and the problem becomes more prominent on small faces and text. To solve this problem, we combine the advantage of MSE-based models and generative models by utilizing region of interest (ROI). We propose Hierarchical-ROI (H-ROI), to split images into several foreground regions and one background region to improve the reconstruction of regions containing faces, text, and complex textures. Further, we propose adaptive quantization by non-linear mapping within the channel dimension to constrain the bit rate while maintaining the visual quality. Exhaustive experiments demonstrate that our methods achieve better visual quality on small faces and text with lower bit rates, e.g., $0.7X$ bits of HiFiC and $0.5X$ bits of BPG.
Authors: Hang Jung Ling, Salom\'e Bru, Julia Puig, Florian Vix\`ege, Simon Mendez, Franck Nicoud, Pierre-Yves Courand, Olivier Bernard, Damien Garcia
Abstract: Intraventricular vector flow mapping (iVFM) seeks to enhance and quantify color Doppler in cardiac imaging. In this study, we propose novel alternatives to the traditional iVFM optimization scheme by utilizing physics-informed neural networks (PINNs) and a physics-guided nnU-Net-based supervised approach. Through rigorous evaluation on simulated color Doppler images derived from a patient-specific computational fluid dynamics model and in vivo Doppler acquisitions, both approaches demonstrate comparable reconstruction performance to the original iVFM algorithm. The efficiency of PINNs is boosted through dual-stage optimization and pre-optimized weights. On the other hand, the nnU-Net method excels in generalizability and real time capabilities. Notably, nnU-Net shows superior robustness on sparse and truncated Doppler data while maintaining independence from explicit boundary conditions. Overall, our results highlight the effectiveness of these methods in reconstructing intraventricular vector blood flow. The study also suggests potential applications of PINNs in ultrafast color Doppler imaging and the incorporation of fluid dynamics equations to derive biomarkers for cardiovascular diseases based on blood flow.
Authors: Divyansh Singhvi, Andrej Erkelens, Raghav Jain, Diganta Misra, Naomi Saphra
Abstract: Measuring nonlinear feature interaction is an established approach to understanding complex patterns of attribution in many models. In this paper, we use Shapley Taylor interaction indices (STII) to analyze the impact of underlying data structure on model representations in a variety of modalities, tasks, and architectures. Considering linguistic structure in masked and auto-regressive language models (MLMs and ALMs), we find that STII increases within idiomatic expressions and that MLMs scale STII with syntactic distance, relying more on syntax in their nonlinear structure than ALMs do. Our speech model findings reflect the phonetic principal that the openness of the oral cavity determines how much a phoneme varies based on its context. Finally, we study image classifiers and illustrate that feature interactions intuitively reflect object boundaries. Our wide range of results illustrates the benefits of interdisciplinary work and domain expertise in interpretability research.
Authors: Aneesh Rangnekar, Nishant Nadkarni, Jue Jiang, Harini Veeraraghavan
Abstract: We assessed the trustworthiness of two self-supervision pretrained transformer models, Swin UNETR and SMIT, for fine-tuned lung (LC) tumor segmentation using 670 CT and MRI scans. We measured segmentation accuracy on two public 3D-CT datasets, robustness on CT scans of patients with COVID-19, CT scans of patients with ovarian cancer and T2-weighted MRI of men with prostate cancer, and zero-shot generalization of LC for T2-weighted MRIs. Both models demonstrated high accuracy on in-distribution data (Dice 0.80 for SMIT and 0.78 for Swin UNETR). SMIT showed similar near-out-of-distribution performance on CT scans (AUROC 89.85% vs. 89.19%) but significantly better far-out-of-distribution accuracy on CT (AUROC 97.2% vs. 87.1%) and MRI (92.15% vs. 73.8%). SMIT outperformed Swin UNETR in zero-shot segmentation on MRI (Dice 0.78 vs. 0.69). We expect these findings to guide the safe development and deployment of current and future pretrained models in routine clinical use.
Authors: Yuexi Du, Regina J. Hooley, John Lewin, Nicha C. Dvornek
Abstract: Digital Breast Tomosynthesis (DBT) is a widely used medical imaging modality for breast cancer screening and diagnosis, offering higher spatial resolution and greater detail through its 3D-like breast volume imaging capability. However, the increased data volume also introduces pronounced data imbalance challenges, where only a small fraction of the volume contains suspicious tissue. This further exacerbates the data imbalance due to the case-level distribution in real-world data and leads to learning a trivial classification model that only predicts the majority class. To address this, we propose a novel method using view-level contrastive Self-supervised Initialization and Fine-Tuning for identifying abnormal DBT images, namely SIFT-DBT. We further introduce a patch-level multi-instance learning method to preserve spatial resolution. The proposed method achieves 92.69% volume-wise AUC on an evaluation of 970 unique studies.
Authors: Isak Samsten, Zed Lee
Abstract: Shapelets are discriminative subsequences, originally embedded in shapelet-based decision trees but have since been extended to shapelet-based transformations. We propose Castor, a simple, efficient, and accurate time series classification algorithm that utilizes shapelets to transform time series. The transformation organizes shapelets into groups with varying dilation and allows the shapelets to compete over the time context to construct a diverse feature representation. By organizing the shapelets into groups, we enable the transformation to transition between levels of competition, resulting in methods that more closely resemble distance-based transformations or dictionary-based transformations. We demonstrate, through an extensive empirical investigation, that Castor yields transformations that result in classifiers that are significantly more accurate than several state-of-the-art classifiers. In an extensive ablation study, we examine the effect of choosing hyperparameters and suggest accurate and efficient default values.
Authors: Masih Eskandar, Tooba Imtiaz, Zifeng Wang, Jennifer Dy
Abstract: The performance of deep models, including Vision Transformers, is known to be vulnerable to adversarial attacks. Many existing defenses against these attacks, such as adversarial training, rely on full-model fine-tuning to induce robustness in the models. These defenses require storing a copy of the entire model, that can have billions of parameters, for each task. At the same time, parameter-efficient prompt tuning is used to adapt large transformer-based models to downstream tasks without the need to save large copies. In this paper, we examine parameter-efficient prompt tuning of Vision Transformers for downstream tasks under the lens of robustness. We show that previous adversarial defense methods, when applied to the prompt tuning paradigm, suffer from gradient obfuscation and are vulnerable to adaptive attacks. We introduce ADAPT, a novel framework for performing adaptive adversarial training in the prompt tuning paradigm. Our method achieves competitive robust accuracy of ~40% w.r.t. SOTA robustness methods using full-model fine-tuning, by tuning only ~1% of the number of parameters.
Authors: Anh Bui, Vy Vo, Tung Pham, Dinh Phung, Trung Le
Abstract: There has long been plenty of theoretical and empirical evidence supporting the success of ensemble learning. Deep ensembles in particular take advantage of training randomness and expressivity of individual neural networks to gain prediction diversity, ultimately leading to better generalization, robustness and uncertainty estimation. In respect of generalization, it is found that pursuing wider local minima result in models being more robust to shifts between training and testing sets. A natural research question arises out of these two approaches as to whether a boost in generalization ability can be achieved if ensemble learning and loss sharpness minimization are integrated. Our work investigates this connection and proposes DASH - a learning algorithm that promotes diversity and flatness within deep ensembles. More concretely, DASH encourages base learners to move divergently towards low-loss regions of minimal sharpness. We provide a theoretical backbone for our method along with extensive empirical evidence demonstrating an improvement in ensemble generalizability.
Authors: Zhenyi Wang, Yan Li, Li Shen, Heng Huang
Abstract: Continual Learning (CL) focuses on learning from dynamic and changing data distributions while retaining previously acquired knowledge. Various methods have been developed to address the challenge of catastrophic forgetting, including regularization-based, Bayesian-based, and memory-replay-based techniques. However, these methods lack a unified framework and common terminology for describing their approaches. This research aims to bridge this gap by introducing a comprehensive and overarching framework that encompasses and reconciles these existing methodologies. Notably, this new framework is capable of encompassing established CL approaches as special instances within a unified and general optimization objective. An intriguing finding is that despite their diverse origins, these methods share common mathematical structures. This observation highlights the compatibility of these seemingly distinct techniques, revealing their interconnectedness through a shared underlying optimization objective. Moreover, the proposed general framework introduces an innovative concept called refresh learning, specifically designed to enhance the CL performance. This novel approach draws inspiration from neuroscience, where the human brain often sheds outdated information to improve the retention of crucial knowledge and facilitate the acquisition of new information. In essence, refresh learning operates by initially unlearning current data and subsequently relearning it. It serves as a versatile plug-in that seamlessly integrates with existing CL methods, offering an adaptable and effective enhancement to the learning process. Extensive experiments on CL benchmarks and theoretical analysis demonstrate the effectiveness of the proposed refresh learning. Code is available at \url{https://github.com/joey-wang123/CL-refresh-learning}.
Authors: Xincheng Yao, Ruoqi Li, Zefeng Qian, Lu Wang, Chongyang Zhang
Abstract: Unified anomaly detection (AD) is one of the most challenges for anomaly detection, where one unified model is trained with normal samples from multiple classes with the objective to detect anomalies in these classes. For such a challenging task, popular normalizing flow (NF) based AD methods may fall into a "homogeneous mapping" issue,where the NF-based AD models are biased to generate similar latent representations for both normal and abnormal features, and thereby lead to a high missing rate of anomalies. In this paper, we propose a novel Hierarchical Gaussian mixture normalizing flow modeling method for accomplishing unified Anomaly Detection, which we call HGAD. Our HGAD consists of two key components: inter-class Gaussian mixture modeling and intra-class mixed class centers learning. Compared to the previous NF-based AD methods, the hierarchical Gaussian mixture modeling approach can bring stronger representation capability to the latent space of normalizing flows, so that even complex multi-class distribution can be well represented and learned in the latent space. In this way, we can avoid mapping different class distributions into the same single Gaussian prior, thus effectively avoiding or mitigating the "homogeneous mapping" issue. We further indicate that the more distinguishable different class centers, the more conducive to avoiding the bias issue. Thus, we further propose a mutual information maximization loss for better structuring the latent feature space. We evaluate our method on four real-world AD benchmarks, where we can significantly improve the previous NF-based AD methods and also outperform the SOTA unified AD methods.
Authors: Qiaojun Yu, Ce Hao, Junbo Wang, Wenhai Liu, Liu Liu, Yao Mu, Yang You, Hengxu Yan, Cewu Lu
Abstract: Robotic manipulation in everyday scenarios, especially in unstructured environments, requires skills in pose-aware object manipulation (POM), which adapts robots' grasping and handling according to an object's 6D pose. Recognizing an object's position and orientation is crucial for effective manipulation. For example, if a mug is lying on its side, it's more effective to grasp it by the rim rather than the handle. Despite its importance, research in POM skills remains limited, because learning manipulation skills requires pose-varying simulation environments and datasets. This paper introduces ManiPose, a pioneering benchmark designed to advance the study of pose-varying manipulation tasks. ManiPose encompasses: 1) Simulation environments for POM feature tasks ranging from 6D pose-specific pick-and-place of single objects to cluttered scenes, further including interactions with articulated objects. 2) A comprehensive dataset featuring geometrically consistent and manipulation-oriented 6D pose labels for 2936 real-world scanned rigid objects and 100 articulated objects across 59 categories. 3) A baseline for POM, leveraging the inferencing abilities of LLM (e.g., ChatGPT) to analyze the relationship between 6D pose and task-specific requirements, offers enhanced pose-aware grasp prediction and motion planning capabilities. Our benchmark demonstrates notable advancements in pose estimation, pose-aware manipulation, and real-robot skill transfer, setting new standards for POM research. We will open-source the ManiPose benchmark with the final version paper, inviting the community to engage with our resources, available at our website:https://sites.google.com/view/manipose.
Authors: Natascha Jeziorski, Claudia Redenbach
Abstract: Training defect detection algorithms for visual surface inspection systems requires a large and representative set of training data. Often there is not enough real data available which additionally cannot cover the variety of possible defects. Synthetic data generated by a synthetic visual surface inspection environment can overcome this problem. Therefore, a digital twin of the object is needed, whose micro-scale surface topography is modeled by texture synthesis models. We develop stochastic texture models for sandblasted and milled surfaces based on topography measurements of such surfaces. As the surface patterns differ significantly, we use separate modeling approaches for the two cases. Sandblasted surfaces are modeled by a combination of data-based texture synthesis methods that rely entirely on the measurements. In contrast, the model for milled surfaces is procedural and includes all process-related parameters known from the machine settings.
Authors: Wenqiao Zhang, Tianwei Lin, Jiang Liu, Fangxun Shu, Haoyuan Li, Lei Zhang, He Wanggui, Hao Zhou, Zheqi Lv, Hao Jiang, Juncheng Li, Siliang Tang, Yueting Zhuang
Abstract: Recent advancements indicate that scaling up Multimodal Large Language Models (MLLMs) effectively enhances performance on downstream multimodal tasks. The prevailing MLLM paradigm, \emph{e.g.}, LLaVA, transforms visual features into text-like tokens using a \emph{static} vision-language mapper, thereby enabling \emph{static} LLMs to develop the capability to comprehend visual information through visual instruction tuning. Although promising, the \emph{static} tuning strategy~\footnote{The static tuning refers to the trained model with static parameters.} that shares the same parameters may constrain performance across different downstream multimodal tasks. In light of this, we introduce HyperLLaVA, which involves adaptive tuning of the projector and LLM parameters, in conjunction with a dynamic visual expert and language expert, respectively. These experts are derived from HyperNetworks, which generates adaptive parameter shifts through visual and language guidance, enabling dynamic projector and LLM modeling in two-stage training. Our experiments demonstrate that our solution significantly surpasses LLaVA on existing MLLM benchmarks, including MME, MMBench, SEED-Bench, and LLaVA-Bench. ~\footnote{Our project is available on the link https://github.com/DCDmllm/HyperLLaVA}.
Authors: Pablo Pueyo, Eduardo Montijano, Ana C. Murillo, Mac Schwager
Abstract: This paper introduces CLIPSwarm, a new algorithm designed to automate the modeling of swarm drone formations based on natural language. The algorithm begins by enriching a provided word, to compose a text prompt that serves as input to an iterative approach to find the formation that best matches the provided word. The algorithm iteratively refines formations of robots to align with the textual description, employing different steps for "exploration" and "exploitation". Our framework is currently evaluated on simple formation targets, limited to contour shapes. A formation is visually represented through alpha-shape contours and the most representative color is automatically found for the input word. To measure the similarity between the description and the visual representation of the formation, we use CLIP [1], encoding text and images into vectors and assessing their similarity. Subsequently, the algorithm rearranges the formation to visually represent the word more effectively, within the given constraints of available drones. Control actions are then assigned to the drones, ensuring robotic behavior and collision-free movement. Experimental results demonstrate the system's efficacy in accurately modeling robot formations from natural language descriptions. The algorithm's versatility is showcased through the execution of drone shows in photorealistic simulation with varying shapes. We refer the reader to the supplementary video for a visual reference of the results.
Authors: Robert Turnbull, Simon Mutch
Abstract: This paper outlines our submission for the 4th COV19D competition as part of the `Domain adaptation, Explainability, Fairness in AI for Medical Image Analysis' (DEF-AI-MIA) workshop at the Computer Vision and Pattern Recognition Conference (CVPR). The competition consists of two challenges. The first is to train a classifier to detect the presence of COVID-19 from over one thousand CT scans from the COV19-CT-DB database. The second challenge is to perform domain adaptation by taking the dataset from Challenge 1 and adding a small number of scans (some annotated and other not) for a different distribution. We preprocessed the CT scans to segment the lungs, and output volumes with the lungs individually and together. We then trained 3D ResNet and Swin Transformer models on these inputs. We annotated the unlabeled CT scans using an ensemble of these models and chose the high-confidence predictions as pseudo-labels for fine-tuning. This resulted in a best cross-validation mean F1 score of 93.39\% for Challenge 1 and a mean F1 score of 92.15 for Challenge 2.
Authors: Kunhang Li, Yansong Feng
Abstract: The task of text2motion is to generate motion sequences from given textual descriptions, where a model should explore the interactions between natural language instructions and human body movements. While most existing works are confined to coarse-grained motion descriptions (e.g., "A man squats."), fine-grained ones specifying movements of relevant body parts are barely explored. Models trained with coarse texts may not be able to learn mappings from fine-grained motion-related words to motion primitives, resulting in the failure in generating motions from unseen descriptions. In this paper, we build a large-scale language-motion dataset with fine-grained textual descriptions, FineHumanML3D, by feeding GPT-3.5-turbo with delicate prompts. Accordingly, we design a new text2motion model, FineMotionDiffuse, which makes full use of fine-grained textual information. Our experiments show that FineMotionDiffuse trained on FineHumanML3D acquires good results in quantitative evaluation. We also find this model can better generate spatially/chronologically composite motions by learning the implicit mappings from simple descriptions to the corresponding basic motions.
Authors: Run He, Huiping Zhuang, Di Fang, Yizhu Chen, Kai Tong, Cen Chen
Abstract: Exemplar-free class-incremental learning (EFCIL) aims to mitigate catastrophic forgetting in class-incremental learning without available historical data. Compared with its counterpart (replay-based CIL) that stores historical samples, the EFCIL suffers more from forgetting issues under the exemplar-free constraint. In this paper, inspired by the recently developed analytic learning (AL) based CIL, we propose a representation enhanced analytic learning (REAL) for EFCIL. The REAL constructs a dual-stream base pretraining (DS-BPT) and a representation enhancing distillation (RED) process to enhance the representation of the extractor. The DS-BPT pretrains model in streams of both supervised learning and self-supervised contrastive learning (SSCL) for base knowledge extraction. The RED process distills the supervised knowledge to the SSCL pretrained backbone and facilitates a subsequent AL-basd CIL that converts the CIL to a recursive least-square problem. Our method addresses the issue of insufficient discriminability in representations of unseen data caused by a frozen backbone in the existing AL-based CIL. Empirical results on various datasets including CIFAR-100, ImageNet-100 and ImageNet-1k, demonstrate that our REAL outperforms the state-of-the-arts in EFCIL, and achieves comparable or even more superior performance compared with the replay-based methods.
Authors: Paloma Garc\'ia-de-Herreros, Vagrant Gautam, Philipp Slusallek, Dietrich Klakow, Marius Mosbach
Abstract: ORCA (Shen et al., 2023) is a recent technique for cross-modal fine-tuning, i.e., applying pre-trained transformer models to modalities beyond their training data. The technique consists primarily of training an embedder and fine-tuning the embedder and model. Despite its high performance on a variety of downstream tasks, we do not understand precisely how each of these components contribute to ORCA's success. Therefore, we run a series of ablations and find that embedder training does not help 2D tasks at all, contrary to what the original paper posits. In 1D tasks, some amount of embedder training is necessary but more is not better. In 4 out of 6 datasets we experiment with, it is model fine-tuning that makes the biggest difference. Through our ablations and baselines, we contribute a better understanding of the individual components of ORCA.
Authors: Mohammod N. I. Suvon, Prasun C. Tripathi, Wenrui Fan, Shuo Zhou, Xianyuan Liu, Samer Alabed, Venet Osmani, Andrew J. Swift, Chen Chen, Haiping Lu
Abstract: Recent advancements in non-invasive detection of cardiac hemodynamic instability (CHDI) primarily focus on applying machine learning techniques to a single data modality, e.g. cardiac magnetic resonance imaging (MRI). Despite their potential, these approaches often fall short especially when the size of labeled patient data is limited, a common challenge in the medical domain. Furthermore, only a few studies have explored multimodal methods to study CHDI, which mostly rely on costly modalities such as cardiac MRI and echocardiogram. In response to these limitations, we propose a novel multimodal variational autoencoder ($\text{CardioVAE}_\text{X,G}$) to integrate low-cost chest X-ray (CXR) and electrocardiogram (ECG) modalities with pre-training on a large unlabeled dataset. Specifically, $\text{CardioVAE}_\text{X,G}$ introduces a novel tri-stream pre-training strategy to learn both shared and modality-specific features, thus enabling fine-tuning with both unimodal and multimodal datasets. We pre-train $\text{CardioVAE}_\text{X,G}$ on a large, unlabeled dataset of $50,982$ subjects from a subset of MIMIC database and then fine-tune the pre-trained model on a labeled dataset of $795$ subjects from the ASPIRE registry. Comprehensive evaluations against existing methods show that $\text{CardioVAE}_\text{X,G}$ offers promising performance (AUROC $=0.79$ and Accuracy $=0.77$), representing a significant step forward in non-invasive prediction of CHDI. Our model also excels in producing fine interpretations of predictions directly associated with clinical features, thereby supporting clinical decision-making.
Authors: Yiwei Lyu, Sung Jik Cha, Cheng Jiang, Asadur Chowdury, Xinhai Hou, Edward Harake, Akhil Kondepudi, Christian Freudiger, Honglak Lee, Todd C. Hollon
Abstract: High-quality, high-resolution medical imaging is essential for clinical care. Raman-based biomedical optical imaging uses non-ionizing infrared radiation to evaluate human tissues in real time and is used for early cancer detection, brain tumor diagnosis, and intraoperative tissue analysis. Unfortunately, optical imaging is vulnerable to image degradation due to laser scattering and absorption, which can result in diagnostic errors and misguided treatment. Restoration of optical images is a challenging computer vision task because the sources of image degradation are multi-factorial, stochastic, and tissue-dependent, preventing a straightforward method to obtain paired low-quality/high-quality data. Here, we present Restorative Step-Calibrated Diffusion (RSCD), an unpaired image restoration method that views the image restoration problem as completing the finishing steps of a diffusion-based image generation task. RSCD uses a step calibrator model to dynamically determine the severity of image degradation and the number of steps required to complete the reverse diffusion process for image restoration. RSCD outperforms other widely used unpaired image restoration methods on both image quality and perceptual evaluation metrics for restoring optical images. Medical imaging experts consistently prefer images restored using RSCD in blinded comparison experiments and report minimal to no hallucinations. Finally, we show that RSCD improves performance on downstream clinical imaging tasks, including automated brain tumor diagnosis and deep tissue imaging. Our code is available at https://github.com/MLNeurosurg/restorative_step-calibrated_diffusion.
URLs: https://github.com/MLNeurosurg/restorative_step-calibrated_diffusion.
Authors: Arun Krishnavajjala, SM Hasan Mansur, Justin Jose, Kevin Moran
Abstract: Recent research has begun to examine the potential of automatically finding and fixing accessibility issues that manifest in software. However, while recent work makes important progress, it has generally been skewed toward identifying issues that affect users with certain disabilities, such as those with visual or hearing impairments. However, there are other groups of users with different types of disabilities that also need software tooling support to improve their experience. As such, this paper aims to automatically identify accessibility issues that affect users with motor-impairments. To move toward this goal, this paper introduces a novel approach, called MotorEase, capable of identifying accessibility issues in mobile app UIs that impact motor-impaired users. Motor-impaired users often have limited ability to interact with touch-based devices, and instead may make use of a switch or other assistive mechanism -- hence UIs must be designed to support both limited touch gestures and the use of assistive devices. MotorEase adapts computer vision and text processing techniques to enable a semantic understanding of app UI screens, enabling the detection of violations related to four popular, previously unexplored UI design guidelines that support motor-impaired users, including: (i) visual touch target size, (ii) expanding sections, (iii) persisting elements, and (iv) adjacent icon visual distance. We evaluate MotorEase on a newly derived benchmark, called MotorCheck, that contains 555 manually annotated examples of violations to the above accessibility guidelines, across 1599 screens collected from 70 applications via a mobile app testing tool. Our experiments illustrate that MotorEase is able to identify violations with an average accuracy of ~90%, and a false positive rate of less than 9%, outperforming baseline techniques.
Authors: Yuxuan Zhou, Xingxing Li, Shengyu Li, Xuanbin Wang, Shaoquan Feng, Yuxuan Tan
Abstract: Visual simultaneous localization and mapping (VSLAM) has broad applications, with state-of-the-art methods leveraging deep neural networks for better robustness and applicability. However, there is a lack of research in fusing these learning-based methods with multi-sensor information, which could be indispensable to push related applications to large-scale and complex scenarios. In this paper, we tightly integrate the trainable deep dense bundle adjustment (DBA) with multi-sensor information through a factor graph. In the framework, recurrent optical flow and DBA are performed among sequential images. The Hessian information derived from DBA is fed into a generic factor graph for multi-sensor fusion, which employs a sliding window and supports probabilistic marginalization. A pipeline for visual-inertial integration is firstly developed, which provides the minimum ability of metric-scale localization and mapping. Furthermore, other sensors (e.g., global navigation satellite system) are integrated for driftless and geo-referencing functionality. Extensive tests are conducted on both public datasets and self-collected datasets. The results validate the superior localization performance of our approach, which enables real-time dense mapping in large-scale environments. The code has been made open-source (https://github.com/GREAT-WHU/DBA-Fusion).
Authors: Dipendra Misra, Akanksha Saran, Tengyang Xie, Alex Lamb, John Langford
Abstract: We study pre-training representations for decision-making using video data, which is abundantly available for tasks such as game agents and software testing. Even though significant empirical advances have been made on this problem, a theoretical understanding remains absent. We initiate the theoretical investigation into principled approaches for representation learning and focus on learning the latent state representations of the underlying MDP using video data. We study two types of settings: one where there is iid noise in the observation, and a more challenging setting where there is also the presence of exogenous noise, which is non-iid noise that is temporally correlated, such as the motion of people or cars in the background. We study three commonly used approaches: autoencoding, temporal contrastive learning, and forward modeling. We prove upper bounds for temporal contrastive learning and forward modeling in the presence of only iid noise. We show that these approaches can learn the latent state and use it to do efficient downstream RL with polynomial sample complexity. When exogenous noise is also present, we establish a lower bound result showing that the sample complexity of learning from video data can be exponentially worse than learning from action-labeled trajectory data. This partially explains why reinforcement learning with video pre-training is hard. We evaluate these representational learning methods in two visual domains, yielding results that are consistent with our theoretical findings.
Authors: Chao Yi, De-Chuan Zhan, Han-Jia Ye
Abstract: Vision Language Models (VLMs) excel in zero-shot image classification by pairing images with textual category names. The expanding variety of Pre-Trained VLMs enhances the likelihood of identifying a suitable VLM for specific tasks. Thus, a promising zero-shot image classification strategy is selecting the most appropriate Pre-Trained VLM from the VLM Zoo, relying solely on the text data of the target dataset without access to the dataset's images. In this paper, we analyze two inherent challenges in assessing the ability of a VLM in this Language-Only VLM selection: the "Modality Gap" -- the disparity in VLM's embeddings across two different modalities, making text a less reliable substitute for images; and the "Capability Gap" -- the discrepancy between the VLM's overall ranking and its ranking for target dataset, hindering direct prediction of a model's dataset-specific performance from its general performance. We propose VLM Selection With gAp Bridging (SWAB) to mitigate the negative impact of these two gaps. SWAB first adopts optimal transport to capture the relevance between open-source datasets and target dataset with a transportation matrix. It then uses this matrix to transfer useful statistics of VLMs from open-source datasets to the target dataset for bridging those two gaps and enhancing the VLM's capacity estimation for VLM selection. Experiments across various VLMs and image classification datasets validate SWAB's effectiveness.
Authors: Xianxu Hou, Linlin Shen, Ke Sun, Guoping Qiu
Abstract: We present a novel method for constructing Variational Autoencoder (VAE). Instead of using pixel-by-pixel loss, we enforce deep feature consistency between the input and the output of a VAE, which ensures the VAE's output to preserve the spatial correlation characteristics of the input, thus leading the output to have a more natural visual appearance and better perceptual quality. Based on recent deep learning works such as style transfer, we employ a pre-trained deep convolutional neural network (CNN) and use its hidden features to define a feature perceptual loss for VAE training. Evaluated on the CelebA face dataset, we show that our model produces better results than other methods in the literature. We also show that our method can produce latent vectors that can capture the semantic information of face expressions and can be used to achieve state-of-the-art performance in facial attribute prediction.
Authors: Zifu Wang, Xuefei Ning, Matthew B. Blaschko
Abstract: Intersection over Union (IoU) losses are surrogates that directly optimize the Jaccard index. Leveraging IoU losses as part of the loss function have demonstrated superior performance in semantic segmentation tasks compared to optimizing pixel-wise losses such as the cross-entropy loss alone. However, we identify a lack of flexibility in these losses to support vital training techniques like label smoothing, knowledge distillation, and semi-supervised learning, mainly due to their inability to process soft labels. To address this, we introduce Jaccard Metric Losses (JMLs), which are identical to the soft Jaccard loss in standard settings with hard labels but are fully compatible with soft labels. We apply JMLs to three prominent use cases of soft labels: label smoothing, knowledge distillation and semi-supervised learning, and demonstrate their potential to enhance model accuracy and calibration. Our experiments show consistent improvements over the cross-entropy loss across 4 semantic segmentation datasets (Cityscapes, PASCAL VOC, ADE20K, DeepGlobe Land) and 13 architectures, including classic CNNs and recent vision transformers. Remarkably, our straightforward approach significantly outperforms state-of-the-art knowledge distillation and semi-supervised learning methods. The code is available at \href{https://github.com/zifuwanggg/JDTLosses}{https://github.com/zifuwanggg/JDTLosses}.
URLs: https://github.com/zifuwanggg/JDTLosses, https://github.com/zifuwanggg/JDTLosses
Authors: Anh-Kiet Duong, Hoang-Lan Nguyen, Toan-Thinh Truong
Abstract: The overall objective of the main project is to propose and develop a system of facial authentication in unlocking phones or applications in phones using facial recognition. The system will include four separate architectures: face detection, face recognition, face spoofing, and classification of closed eyes. In which, we consider the problem of face recognition to be the most important, determining the true identity of the person standing in front of the screen with absolute accuracy is what facial recognition systems need to achieve. Along with the development of the face recognition problem, the problem of the anti-fake face is also gradually becoming popular and equally important. Our goal is to propose and develop two loss functions: LMCot and Double Loss. Then apply them to the face authentication process.
Authors: Juil Koo, Seungwoo Yoo, Minh Hieu Nguyen, Minhyuk Sung
Abstract: We present a cascaded diffusion model based on a part-level implicit 3D representation. Our model achieves state-of-the-art generation quality and also enables part-level shape editing and manipulation without any additional training in conditional setup. Diffusion models have demonstrated impressive capabilities in data generation as well as zero-shot completion and editing via a guided reverse process. Recent research on 3D diffusion models has focused on improving their generation capabilities with various data representations, while the absence of structural information has limited their capability in completion and editing tasks. We thus propose our novel diffusion model using a part-level implicit representation. To effectively learn diffusion with high-dimensional embedding vectors of parts, we propose a cascaded framework, learning diffusion first on a low-dimensional subspace encoding extrinsic parameters of parts and then on the other high-dimensional subspace encoding intrinsic attributes. In the experiments, we demonstrate the outperformance of our method compared with the previous ones both in generation and part-level completion and manipulation tasks.
Authors: Nathanael L. Baisa
Abstract: In this paper, we propose a multi-task representation learning framework to jointly estimate the identity, gender and age of individuals from their hand images for the purpose of criminal investigations since the hand images are often the only available information in cases of serious crime such as sexual abuse. We investigate different up-to-date deep learning architectures and compare their performance for joint estimation of identity, gender and age from hand images of perpetrators of serious crime. To simplify the age prediction, we create age groups for the age estimation. We make extensive evaluations and comparisons of both convolution-based and transformer-based deep learning architectures on a publicly available 11k hands dataset. Our experimental analysis shows that it is possible to efficiently estimate not only identity but also other attributes such as gender and age of suspects jointly from hand images for criminal investigations, which is crucial in assisting international police forces in the court to identify and convict abusers.
Authors: Zifu Wang, Teodora Popordanoska, Jeroen Bertels, Robin Lemmens, Matthew B. Blaschko
Abstract: The soft Dice loss (SDL) has taken a pivotal role in numerous automated segmentation pipelines in the medical imaging community. Over the last years, some reasons behind its superior functioning have been uncovered and further optimizations have been explored. However, there is currently no implementation that supports its direct utilization in scenarios involving soft labels. Hence, a synergy between the use of SDL and research leveraging the use of soft labels, also in the context of model calibration, is still missing. In this work, we introduce Dice semimetric losses (DMLs), which (i) are by design identical to SDL in a standard setting with hard labels, but (ii) can be employed in settings with soft labels. Our experiments on the public QUBIQ, LiTS and KiTS benchmarks confirm the potential synergy of DMLs with soft labels (e.g. averaging, label smoothing, and knowledge distillation) over hard labels (e.g. majority voting and random selection). As a result, we obtain superior Dice scores and model calibration, which supports the wider adoption of DMLs in practice. The code is available at https://github.com/zifuwanggg/JDTLosses
Authors: Yuang Ai, Xiaoqiang Zhou, Huaibo Huang, Lei Zhang, Ran He
Abstract: Unsupervised Domain Adaptation (UDA) can effectively address domain gap issues in real-world image Super-Resolution (SR) by accessing both the source and target data. Considering privacy policies or transmission restrictions of source data in practical scenarios, we propose a SOurce-free Domain Adaptation framework for image SR (SODA-SR) to address this issue, i.e., adapt a source-trained model to a target domain with only unlabeled target data. SODA-SR leverages the source-trained model to generate refined pseudo-labels for teacher-student learning. To better utilize pseudo-labels, we propose a novel wavelet-based augmentation method, named Wavelet Augmentation Transformer (WAT), which can be flexibly incorporated with existing networks, to implicitly produce useful augmented data. WAT learns low-frequency information of varying levels across diverse samples, which is aggregated efficiently via deformable attention. Furthermore, an uncertainty-aware self-training mechanism is proposed to improve the accuracy of pseudo-labels, with inaccurate predictions being rectified by uncertainty estimation. To acquire better SR results and avoid overfitting pseudo-labels, several regularization losses are proposed to constrain target LR and SR images in the frequency domain. Experiments show that without accessing source data, SODA-SR outperforms state-of-the-art UDA methods in both synthetic$\rightarrow$real and real$\rightarrow$real adaptation settings, and is not constrained by specific network architectures.
Authors: Rustam Tagiew, Martin K\"oppel, Karsten Schwalbe, Patrick Denzler, Philipp Neumaier, Tobias Klockau, Martin Boekhoff, Pavel Klasek, Roman Tilly
Abstract: To achieve a driverless train operation on mainline railways, actual and potential obstacles for the train's driveway must be detected automatically by appropriate sensor systems. Machine learning algorithms have proven to be powerful tools for this task during the last years. However, these algorithms require large amounts of high-quality annotated data containing railway-specific objects as training data. Unfortunately, all of the publicly available datasets that tackle this requirement are restricted in some way. Therefore, this paper presents OSDaR23, a multi-sensor dataset of 45 subsequences acquired in Hamburg, Germany, in September 2021, that was created to foster driverless train operation on mainline railways. The sensor setup consists of multiple calibrated and synchronized infrared (IR) and visual (RGB) cameras, lidars, a radar, and position and acceleration sensors mounted on the front of a rail vehicle. In addition to the raw data, the dataset contains 204091 polyline, polygonal, rectangle, and cuboid annotations in total for 20 different object classes. It is the first publicly available multi-sensor dataset annotated with a variety of object classes that are relevant for the railway context. OSDaR23, available at data.fid-move.de/dataset/osdar23, can also be used for tasks beyond collision prediction, which are listed in this paper.
Authors: Runqi Wang, Xiaoyue Duan, Guoliang Kang, Jianzhuang Liu, Shaohui Lin, Songcen Xu, Jinhu Lv, Baochang Zhang
Abstract: Continual learning aims to enable a model to incrementally learn knowledge from sequentially arrived data. Previous works adopt the conventional classification architecture, which consists of a feature extractor and a classifier. The feature extractor is shared across sequentially arrived tasks or classes, but one specific group of weights of the classifier corresponding to one new class should be incrementally expanded. Consequently, the parameters of a continual learner gradually increase. Moreover, as the classifier contains all historical arrived classes, a certain size of the memory is usually required to store rehearsal data to mitigate classifier bias and catastrophic forgetting. In this paper, we propose a non-incremental learner, named AttriCLIP, to incrementally extract knowledge of new classes or tasks. Specifically, AttriCLIP is built upon the pre-trained visual-language model CLIP. Its image encoder and text encoder are fixed to extract features from both images and text. Text consists of a category name and a fixed number of learnable parameters which are selected from our designed attribute word bank and serve as attributes. As we compute the visual and textual similarity for classification, AttriCLIP is a non-incremental learner. The attribute prompts, which encode the common knowledge useful for classification, can effectively mitigate the catastrophic forgetting and avoid constructing a replay memory. We evaluate our AttriCLIP and compare it with CLIP-based and previous state-of-the-art continual learning methods in realistic settings with domain-shift and long-sequence learning. The results show that our method performs favorably against previous state-of-the-arts. The implementation code can be available at https://github.com/bhrqw/AttriCLIP.
Authors: Takato Yoshikawa, Yuki Endo, Yoshihiro Kanamori
Abstract: This paper tackles text-guided control of StyleGAN for editing garments in full-body human images. Existing StyleGAN-based methods suffer from handling the rich diversity of garments and body shapes and poses. We propose a framework for text-guided full-body human image synthesis via an attention-based latent code mapper, which enables more disentangled control of StyleGAN than existing mappers. Our latent code mapper adopts an attention mechanism that adaptively manipulates individual latent codes on different StyleGAN layers under text guidance. In addition, we introduce feature-space masking at inference time to avoid unwanted changes caused by text inputs. Our quantitative and qualitative evaluations reveal that our method can control generated images more faithfully to given texts than existing methods.
Authors: Fuxiao Liu, Kevin Lin, Linjie Li, Jianfeng Wang, Yaser Yacoob, Lijuan Wang
Abstract: Despite the promising progress in multi-modal tasks, current large multi-modal models (LMMs) are prone to hallucinating inconsistent descriptions with respect to the associated image and human instructions. This paper addresses this issue by introducing the first large and diverse visual instruction tuning dataset, named Large-scale Robust Visual (LRV)-Instruction. Our dataset comprises 400k visual instructions generated by GPT4, covering 16 vision-and-language tasks with open-ended instructions and answers. Unlike existing studies that primarily focus on positive instruction samples, we design LRV-Instruction to include both positive and negative instructions for more robust visual instruction tuning. Our negative instructions are designed at three semantic levels: (i) Nonexistent Object Manipulation, (ii) Existent Object Manipulation and (iii) Knowledge Manipulation. To efficiently measure the hallucination generated by LMMs, we propose GPT4-Assisted Visual Instruction Evaluation (GAVIE), a stable approach to evaluate visual instruction tuning like human experts. GAVIE does not require human-annotated groundtruth answers and can adapt to diverse instruction formats. We conduct comprehensive experiments to investigate the hallucination of LMMs. Our results demonstrate existing LMMs exhibit significant hallucinations when presented with our negative instructions, particularly Existent Object and Knowledge Manipulation instructions. Moreover, we successfully mitigate hallucination by finetuning MiniGPT4 and mPLUG-Owl on LRV-Instruction while improving performance on several public datasets compared to state-of-the-art methods. Additionally, we observed that a balanced ratio of positive and negative instances in the training data leads to a more robust model. Code and data are available at https://github.com/FuxiaoLiu/LRV-Instruction.
Authors: Jiachun Pan, Jun Hao Liew, Vincent Y. F. Tan, Jiashi Feng, Hanshu Yan
Abstract: Existing customization methods require access to multiple reference examples to align pre-trained diffusion probabilistic models (DPMs) with user-provided concepts. This paper aims to address the challenge of DPM customization when the only available supervision is a differentiable metric defined on the generated contents. Since the sampling procedure of DPMs involves recursive calls to the denoising UNet, na\"ive gradient backpropagation requires storing the intermediate states of all iterations, resulting in extremely high memory consumption. To overcome this issue, we propose a novel method AdjointDPM, which first generates new samples from diffusion models by solving the corresponding probability-flow ODEs. It then uses the adjoint sensitivity method to backpropagate the gradients of the loss to the models' parameters (including conditioning signals, network weights, and initial noises) by solving another augmented ODE. To reduce numerical errors in both the forward generation and gradient backpropagation processes, we further reparameterize the probability-flow ODE and augmented ODE as simple non-stiff ODEs using exponential integration. Finally, we demonstrate the effectiveness of AdjointDPM on three interesting tasks: converting visual effects into identification text embeddings, finetuning DPMs for specific types of stylization, and optimizing initial noise to generate adversarial samples for security auditing.
Authors: Ye Tian, Mengyu Yang, Lanshan Zhang, Zhizhen Zhang, Yang Liu, Xiaohui Xie, Xirong Que, Wendong Wang
Abstract: Recent adaptive methods for efficient video recognition mostly follow the two-stage paradigm of "preview-then-recognition" and have achieved great success on multiple video benchmarks. However, this two-stage paradigm involves two visits of raw frames from coarse-grained to fine-grained during inference (cannot be parallelized), and the captured spatiotemporal features cannot be reused in the second stage (due to varying granularity), being not friendly to efficiency and computation optimization. To this end, inspired by human cognition, we propose a novel recognition paradigm of "View while Moving" for efficient long-untrimmed video recognition. In contrast to the two-stage paradigm, our paradigm only needs to access the raw frame once. The two phases of coarse-grained sampling and fine-grained recognition are combined into unified spatiotemporal modeling, showing great performance. Moreover, we investigate the properties of semantic units in video and propose a hierarchical mechanism to efficiently capture and reason about the unit-level and video-level temporal semantics in long-untrimmed videos respectively. Extensive experiments on both long-untrimmed and short-trimmed videos demonstrate that our approach outperforms state-of-the-art methods in terms of accuracy as well as efficiency, yielding new efficiency and accuracy trade-offs for video spatiotemporal modeling.
Authors: Weiwen Chen, Yingtie Lei, Shenghong Luo, Xuhang Chen, Ziyang Zhou, Mingxian Li, Chi-Man Pun
Abstract: Document shadow is a common issue that arises when capturing documents using mobile devices, which significantly impacts readability. Current methods encounter various challenges, including inaccurate detection of shadow masks and estimation of illumination. In this paper, we propose ShaDocFormer, a Transformer-based architecture that integrates traditional methodologies and deep learning techniques to tackle the problem of document shadow removal. The ShaDocFormer architecture comprises two components: the Shadow-attentive Threshold Detector (STD) and the Cascaded Fusion Refiner (CFR). The STD module employs a traditional thresholding technique and leverages the attention mechanism of the Transformer to gather global information, thereby enabling precise detection of shadow masks. The cascaded and aggregative structure of the CFR module facilitates a coarse-to-fine restoration process for the entire image. As a result, ShaDocFormer excels in accurately detecting and capturing variations in both shadow and illumination, thereby enabling effective removal of shadows. Extensive experiments demonstrate that ShaDocFormer outperforms current state-of-the-art methods in both qualitative and quantitative measurements.
Authors: Roberto Pellerito, Marco Cannici, Daniel Gehrig, Joris Belhadj, Olivier Dubois-Matra, Massimo Casasco, Davide Scaramuzza
Abstract: Visual Odometry (VO) is crucial for autonomous robotic navigation, especially in GPS-denied environments like planetary terrains. To improve robustness, recent model-based VO systems have begun combining standard and event-based cameras. Event cameras excel in low-light and high-speed motion, while standard cameras provide dense and easier-to-track features, even in low-textured areas. However, the field of image- and event-based VO still predominantly relies on model-based methods and is yet to fully integrate recent image-only advancements leveraging end-to-end learning-based architectures. Seamlessly integrating the two modalities remains challenging due to their different nature, one asynchronous, the other not, limiting the potential for a more effective image- and event-based VO. We introduce RAMP-VO, the first end-to-end learned image- and event-based VO system. It leverages novel Recurrent, Asynchronous, and Massively Parallel (RAMP) encoders capable of fusing asynchronous events with image data, providing 8x faster inference and 33% more accurate predictions than existing solutions. Despite being trained only in simulation, RAMP-VO outperforms image- and event-based methods by 46% and 60%, respectively, on traditional, real-world benchmarks as well as newly introduced Apollo and Malapert landing sequences, paving the way for robust and asynchronous VO in space.
Authors: Yingtie Lei, Weiwen Chen, Shenghong Luo, Ziyang Zhou, Mingxian Li, Chi-Man Pun
Abstract: Underwater images often exhibit poor quality, distorted color balance and low contrast due to the complex and intricate interplay of light, water, and objects. Despite the significant contributions of previous underwater enhancement techniques, there exist several problems that demand further improvement: (i) The current deep learning methods rely on Convolutional Neural Networks (CNNs) that lack the multi-scale enhancement, and global perception field is also limited. (ii) The scarcity of paired real-world underwater datasets poses a significant challenge, and the utilization of synthetic image pairs could lead to overfitting. To address the aforementioned problems, this paper introduces a Multi-scale Transformer-based Network called UWFormer for enhancing images at multiple frequencies via semi-supervised learning, in which we propose a Nonlinear Frequency-aware Attention mechanism and a Multi-Scale Fusion Feed-forward Network for low-frequency enhancement. Besides, we introduce a special underwater semi-supervised training strategy, where we propose a Subaqueous Perceptual Loss function to generate reliable pseudo labels. Experiments using full-reference and non-reference underwater benchmarks demonstrate that our method outperforms state-of-the-art methods in terms of both quantity and visual quality.
Authors: Jianyuan Zhang, Zhiliu Yang, Meng Zhang
Abstract: Large-scale semantic mapping is crucial for outdoor autonomous agents to fulfill high-level tasks such as planning and navigation. This paper proposes a novel method for large-scale 3D semantic reconstruction through implicit representations from posed LiDAR measurements alone. We first leverage an octree-based and hierarchical structure to store implicit features, then these implicit features are decoded to semantic information and signed distance value through shallow Multilayer Perceptrons (MLPs). We adopt off-the-shelf algorithms to predict the semantic labels and instance IDs of point clouds. We then jointly optimize the feature embeddings and MLPs parameters with a self-supervision paradigm for point cloud geometry and a pseudo-supervision paradigm for semantic and panoptic labels. Subsequently, categories and geometric structures for novel points are regressed, and marching cubes are exploited to subdivide and visualize the scenes in the inferring stage. For scenarios with memory constraints, a map stitching strategy is also developed to merge sub-maps into a complete map. Experiments on two real-world datasets, SemanticKITTI and SemanticPOSS, demonstrate the superior segmentation efficiency and mapping effectiveness of our framework compared to current state-of-the-art 3D LiDAR mapping methods.
Authors: Juil Koo, Chanho Park, Minhyuk Sung
Abstract: We introduce Posterior Distillation Sampling (PDS), a novel optimization method for parametric image editing based on diffusion models. Existing optimization-based methods, which leverage the powerful 2D prior of diffusion models to handle various parametric images, have mainly focused on generation. Unlike generation, editing requires a balance between conforming to the target attribute and preserving the identity of the source content. Recent 2D image editing methods have achieved this balance by leveraging the stochastic latent encoded in the generative process of diffusion models. To extend the editing capabilities of diffusion models shown in pixel space to parameter space, we reformulate the 2D image editing method into an optimization form named PDS. PDS matches the stochastic latents of the source and the target, enabling the sampling of targets in diverse parameter spaces that align with a desired attribute while maintaining the source's identity. We demonstrate that this optimization resembles running a generative process with the target attribute, but aligning this process with the trajectory of the source's generative process. Extensive editing results in Neural Radiance Fields and Scalable Vector Graphics representations demonstrate that PDS is capable of sampling targets to fulfill the aforementioned balance across various parameter spaces.
Authors: Yurui Chen, Chun Gu, Junzhe Jiang, Xiatian Zhu, Li Zhang
Abstract: Modeling dynamic, large-scale urban scenes is challenging due to their highly intricate geometric structures and unconstrained dynamics in both space and time. Prior methods often employ high-level architectural priors, separating static and dynamic elements, resulting in suboptimal capture of their synergistic interactions. To address this challenge, we present a unified representation model, called Periodic Vibration Gaussian (PVG). PVG builds upon the efficient 3D Gaussian splatting technique, originally designed for static scene representation, by introducing periodic vibration-based temporal dynamics. This innovation enables PVG to elegantly and uniformly represent the characteristics of various objects and elements in dynamic urban scenes. To enhance temporally coherent and large scene representation learning with sparse training data, we introduce a novel temporal smoothing mechanism and a position-aware adaptive control strategy respectively. Extensive experiments on Waymo Open Dataset and KITTI benchmarks demonstrate that PVG surpasses state-of-the-art alternatives in both reconstruction and novel view synthesis for both dynamic and static scenes. Notably, PVG achieves this without relying on manually labeled object bounding boxes or expensive optical flow estimation. Moreover, PVG exhibits 900-fold acceleration in rendering over the best alternative.
Authors: Hai Zhang, Junzhe Xu, Shanlin Jiang, Zhenan He
Abstract: Learning from a limited amount of data, namely Few-Shot Learning, stands out as a challenging computer vision task. Several works exploit semantics and design complicated semantic fusion mechanisms to compensate for rare representative features within restricted data. However, relying on naive semantics such as class names introduces biases due to their brevity, while acquiring extensive semantics from external knowledge takes a huge time and effort. This limitation severely constrains the potential of semantics in few-shot learning. In this paper, we design an automatic way called Semantic Evolution to generate high-quality semantics. The incorporation of high-quality semantics alleviates the need for complex network structures and learning algorithms used in previous works. Hence, we employ a simple two-layer network termed Semantic Alignment Network to transform semantics and visual features into robust class prototypes with rich discriminative features for few-shot classification. The experimental results show our framework outperforms all previous methods on six benchmarks, demonstrating a simple network with high-quality semantics can beat intricate multi-modal modules on few-shot classification tasks. Code is available at https://github.com/zhangdoudou123/SemFew.
Authors: Jinxin Zhou, Tianyu Ding, Tianyi Chen, Jiachen Jiang, Ilya Zharkov, Zhihui Zhu, Luming Liang
Abstract: We present DREAM, a novel training framework representing Diffusion Rectification and Estimation Adaptive Models, requiring minimal code changes (just three lines) yet significantly enhancing the alignment of training with sampling in diffusion models. DREAM features two components: diffusion rectification, which adjusts training to reflect the sampling process, and estimation adaptation, which balances perception against distortion. When applied to image super-resolution (SR), DREAM adeptly navigates the tradeoff between minimizing distortion and preserving high image quality. Experiments demonstrate DREAM's superiority over standard diffusion-based SR methods, showing a $2$ to $3\times $ faster training convergence and a $10$ to $20\times$ reduction in sampling steps to achieve comparable results. We hope DREAM will inspire a rethinking of diffusion model training paradigms.
Authors: Pengxiang Li, Kai Chen, Zhili Liu, Ruiyuan Gao, Lanqing Hong, Guo Zhou, Hua Yao, Dit-Yan Yeung, Huchuan Lu, Xu Jia
Abstract: Despite remarkable achievements in video synthesis, achieving granular control over complex dynamics, such as nuanced movement among multiple interacting objects, still presents a significant hurdle for dynamic world modeling, compounded by the necessity to manage appearance and disappearance, drastic scale changes, and ensure consistency for instances across frames. These challenges hinder the development of video generation that can faithfully mimic real-world complexity, limiting utility for applications requiring high-level realism and controllability, including advanced scene simulation and training of perception systems. To address that, we propose TrackDiffusion, a novel video generation framework affording fine-grained trajectory-conditioned motion control via diffusion models, which facilitates the precise manipulation of the object trajectories and interactions, overcoming the prevalent limitation of scale and continuity disruptions. A pivotal component of TrackDiffusion is the instance enhancer, which explicitly ensures inter-frame consistency of multiple objects, a critical factor overlooked in the current literature. Moreover, we demonstrate that generated video sequences by our TrackDiffusion can be used as training data for visual perception models. To the best of our knowledge, this is the first work to apply video diffusion models with tracklet conditions and demonstrate that generated frames can be beneficial for improving the performance of object trackers.
Authors: Qiang Wen, Yazhou Xing, Zhefan Rao, Qifeng Chen
Abstract: Enhancing a low-light noisy RAW image into a well-exposed and clean sRGB image is a significant challenge for modern digital cameras. Prior approaches have difficulties in recovering fine-grained details and true colors of the scene under extremely low-light environments due to near-to-zero SNR. Meanwhile, diffusion models have shown significant progress towards general domain image generation. In this paper, we propose to leverage the pre-trained latent diffusion model to perform the neural ISP for enhancing extremely low-light images. Specifically, to tailor the pre-trained latent diffusion model to operate on the RAW domain, we train a set of lightweight taming modules to inject the RAW information into the diffusion denoising process via modulating the intermediate features of UNet. We further observe different roles of UNet denoising and decoder reconstruction in the latent diffusion model, which inspires us to decompose the low-light image enhancement task into latent-space low-frequency content generation and decoding-phase high-frequency detail maintenance. Through extensive experiments on representative datasets, we demonstrate our simple design not only achieves state-of-the-art performance in quantitative evaluations but also shows significant superiority in visual comparisons over strong baselines, which highlight the effectiveness of powerful generative priors for neural ISP under extremely low-light environments. The project page is available at https://csqiangwen.github.io/projects/ldm-isp/
Authors: Niki Amini-Naieni, Tomas Jakab, Andrea Vedaldi, Ronald Clark
Abstract: Although Neural Radiance Fields (NeRFs) have markedly improved novel view synthesis, accurate uncertainty quantification in their image predictions remains an open problem. The prevailing methods for estimating uncertainty, including the state-of-the-art Density-aware NeRF Ensembles (DANE) [29], quantify uncertainty without calibration. This frequently leads to over- or under-confidence in image predictions, which can undermine their real-world applications. In this paper, we propose a method which, for the first time, achieves calibrated uncertainties for NeRFs. To accomplish this, we overcome a significant challenge in adapting existing calibration techniques to NeRFs: a need to hold out ground truth images from the target scene, reducing the number of images left to train the NeRF. This issue is particularly problematic in sparse-view settings, where we can operate with as few as three images. To address this, we introduce the concept of a meta-calibrator that performs uncertainty calibration for NeRFs with a single forward pass without the need for holding out any images from the target scene. Our meta-calibrator is a neural network that takes as input the NeRF images and uncalibrated uncertainty maps and outputs a scene-specific calibration curve that corrects the NeRF's uncalibrated uncertainties. We show that the meta-calibrator can generalize on unseen scenes and achieves well-calibrated and state-of-the-art uncertainty for NeRFs, significantly beating DANE and other approaches. This opens opportunities to improve applications that rely on accurate NeRF uncertainty estimates such as next-best view planning and potentially more trustworthy image reconstruction for medical diagnosis.
Authors: Tero Karras, Miika Aittala, Jaakko Lehtinen, Janne Hellsten, Timo Aila, Samuli Laine
Abstract: Diffusion models currently dominate the field of data-driven image synthesis with their unparalleled scaling to large datasets. In this paper, we identify and rectify several causes for uneven and ineffective training in the popular ADM diffusion model architecture, without altering its high-level structure. Observing uncontrolled magnitude changes and imbalances in both the network activations and weights over the course of training, we redesign the network layers to preserve activation, weight, and update magnitudes on expectation. We find that systematic application of this philosophy eliminates the observed drifts and imbalances, resulting in considerably better networks at equal computational complexity. Our modifications improve the previous record FID of 2.41 in ImageNet-512 synthesis to 1.81, achieved using fast deterministic sampling. As an independent contribution, we present a method for setting the exponential moving average (EMA) parameters post-hoc, i.e., after completing the training run. This allows precise tuning of EMA length without the cost of performing several training runs, and reveals its surprising interactions with network architecture, training time, and guidance.
Authors: Yuang Ai, Huaibo Huang, Xiaoqiang Zhou, Jiexiang Wang, Ran He
Abstract: Despite substantial progress, all-in-one image restoration (IR) grapples with persistent challenges in handling intricate real-world degradations. This paper introduces MPerceiver: a novel multimodal prompt learning approach that harnesses Stable Diffusion (SD) priors to enhance adaptiveness, generalizability and fidelity for all-in-one image restoration. Specifically, we develop a dual-branch module to master two types of SD prompts: textual for holistic representation and visual for multiscale detail representation. Both prompts are dynamically adjusted by degradation predictions from the CLIP image encoder, enabling adaptive responses to diverse unknown degradations. Moreover, a plug-in detail refinement module improves restoration fidelity via direct encoder-to-decoder information transformation. To assess our method, MPerceiver is trained on 9 tasks for all-in-one IR and outperforms state-of-the-art task-specific methods across most tasks. Post multitask pre-training, MPerceiver attains a generalized representation in low-level vision, exhibiting remarkable zero-shot and few-shot capabilities in unseen tasks. Extensive experiments on 16 IR tasks underscore the superiority of MPerceiver in terms of adaptiveness, generalizability and fidelity.
Authors: Genki Kinoshita, Ko Nishino
Abstract: In this paper, we introduce a novel training method for making any monocular depth network learn absolute scale and estimate metric road-scene depth just from regular training data, i.e., driving videos. We refer to this training framework as StableCamH. The key idea is to leverage cars found on the road as sources of scale supervision but to incorporate them in the training robustly. StableCamH detects and estimates the sizes of cars in the frame and aggregates scale information extracted from them into a camera height estimate whose consistency across the entire video sequence is enforced as scale supervision. This realizes robust unsupervised training of any, otherwise scale-oblivious, monocular depth network to become not only scale-aware but also metric-accurate without the need for auxiliary sensors and extra supervision. Extensive experiments on the KITTI and Cityscapes datasets show the effectiveness of StableCamH and its state-of-the-art accuracy compared with related methods. We also show that StableCamH enables training on mixed datasets of different camera heights, which leads to larger-scale training and thus higher generalization. Metric depth reconstruction is essential in any road-scene visual modeling, and StableCamH democratizes its deployment by establishing the means to train any model as a metric depth estimator.
Authors: Osman \"Ulger, Maksymilian Kulicki, Yuki Asano, Martin R. Oswald
Abstract: Open-ended image understanding tasks gained significant attention from the research community, particularly with the emergence of Vision-Language Models. Open-Vocabulary Segmentation (OVS) methods are capable of performing semantic segmentation without relying on a fixed vocabulary, and in some cases, they operate without the need for training or fine-tuning. However, OVS methods typically require users to specify the vocabulary based on the task or dataset at hand. In this paper, we introduce \textit{Auto-Vocabulary Semantic Segmentation (AVS)}, advancing open-ended image understanding by eliminating the necessity to predefine object categories for segmentation. Our approach, \ours, presents a framework that autonomously identifies relevant class names using enhanced BLIP embeddings, which are utilized for segmentation afterwards. Given that open-ended object category predictions cannot be directly compared with a fixed ground truth, we develop a Large Language Model-based Auto-Vocabulary Evaluator (LAVE) to efficiently evaluate the automatically generated class names and their corresponding segments. Our method sets new benchmarks on datasets such as PASCAL VOC and Context, ADE20K, and Cityscapes for AVS and showcases competitive performance to OVS methods that require specified class names.
Authors: Prakhar Srivastava, Ruihan Yang, Gavin Kerrigan, Gideon Dresdner, Jeremy McGibbon, Christopher Bretherton, Stephan Mandt
Abstract: In climate science and meteorology, high-resolution local precipitation (rain and snowfall) predictions are limited by the computational costs of simulation-based methods. Statistical downscaling, or super-resolution, is a common workaround where a low-resolution prediction is improved using statistical approaches. Unlike traditional computer vision tasks, weather and climate applications require capturing the accurate conditional distribution of high-resolution given low-resolution patterns to assure reliable ensemble averages and unbiased estimates of extreme events, such as heavy rain. This work extends recent video diffusion models to precipitation super-resolution, employing a deterministic downscaler followed by a temporally-conditioned diffusion model to capture noise characteristics and high-frequency patterns. We test our approach on FV3GFS output, an established large-scale global atmosphere model, and compare it against five state-of-the-art baselines. Our analysis, capturing CRPS, MSE, precipitation distributions, and qualitative aspects using California and the Himalayas as examples, establishes our method as a new standard for data-driven precipitation downscaling.
Authors: Rao Fu, Zehao Wen, Zichen Liu, Srinath Sridhar
Abstract: Inspired by cognitive theories, we introduce AnyHome, a framework that translates any text into well-structured and textured indoor scenes at a house-scale. By prompting Large Language Models (LLMs) with designed templates, our approach converts provided textual narratives into amodal structured representations. These representations guarantee consistent and realistic spatial layouts by directing the synthesis of a geometry mesh within defined constraints. A Score Distillation Sampling process is then employed to refine the geometry, followed by an egocentric inpainting process that adds lifelike textures to it. AnyHome stands out with its editability, customizability, diversity, and realism. The structured representations for scenes allow for extensive editing at varying levels of granularity. Capable of interpreting texts ranging from simple labels to detailed narratives, AnyHome generates detailed geometries and textures that outperform existing methods in both quantitative and qualitative measures.
Authors: Henry Hengyuan Zhao, Pan Zhou, Mike Zheng Shou
Abstract: Instruction tuning data is essential for training the Multimodal Large Language Models (MLLMs). However, the creation of high-quality instruction tuning data presents significant challenges. Prior methods that depended on GPT-4 for data generation were not only costly but also lacked satisfactory performance in complex tasks (i.e., grounding-based reasoning tasks). To address these issues, we developed an innovative data generation pipeline, Genixer, to generate various high-quality instruction tuning data, including nine representative tasks, e.g., Common VQA, REC, REG, and PointQ. Specifically, Genixer provides a unified solution with four key steps for alleviating the difficulty of data generation: (i) instruction data collection, (ii) instruction template design, (iii) empowering MLLM, and (iv) data generation and filtering. Subsequently, the superior qualitative results of our Genixer demonstrate that current MLLMs have a strong potential to evolve into powerful data generators. Additionally, to validate the efficacy of generated data quantitatively, we add the instruction tuning data produced by Genixer into the training of two representative MLLMs and observe the consistent improvements on various VQA tasks and multimodal benchmarks.
Authors: Ayush Singh, Aayush J Rana, Akash Kumar, Shruti Vyas, Yogesh Singh Rawat
Abstract: In this work, we focus on label efficient learning for video action detection. We develop a novel semi-supervised active learning approach which utilizes both labeled as well as unlabeled data along with informative sample selection for action detection. Video action detection requires spatio-temporal localization along with classification, which poses several challenges for both active learning informative sample selection as well as semi-supervised learning pseudo label generation. First, we propose NoiseAug, a simple augmentation strategy which effectively selects informative samples for video action detection. Next, we propose fft-attention, a novel technique based on high-pass filtering which enables effective utilization of pseudo label for SSL in video action detection by emphasizing on relevant activity region within a video. We evaluate the proposed approach on three different benchmark datasets, UCF-101-24, JHMDB-21, and Youtube-VOS. First, we demonstrate its effectiveness on video action detection where the proposed approach outperforms prior works in semi-supervised and weakly-supervised learning along with several baseline approaches in both UCF101-24 and JHMDB-21. Next, we also show its effectiveness on Youtube-VOS for video object segmentation demonstrating its generalization capability for other dense prediction tasks in videos.
Authors: Xiaoyu Zhou, Zhiwei Lin, Xiaojun Shan, Yongtao Wang, Deqing Sun, Ming-Hsuan Yang
Abstract: We present DrivingGaussian, an efficient and effective framework for surrounding dynamic autonomous driving scenes. For complex scenes with moving objects, we first sequentially and progressively model the static background of the entire scene with incremental static 3D Gaussians. We then leverage a composite dynamic Gaussian graph to handle multiple moving objects, individually reconstructing each object and restoring their accurate positions and occlusion relationships within the scene. We further use a LiDAR prior for Gaussian Splatting to reconstruct scenes with greater details and maintain panoramic consistency. DrivingGaussian outperforms existing methods in dynamic driving scene reconstruction and enables photorealistic surround-view synthesis with high-fidelity and multi-camera consistency. Our project page is at: https://github.com/VDIGPKU/DrivingGaussian.
Authors: Jiwoo Chung, Sangeek Hyun, Jae-Pil Heo
Abstract: Despite the impressive generative capabilities of diffusion models, existing diffusion model-based style transfer methods require inference-stage optimization (e.g. fine-tuning or textual inversion of style) which is time-consuming, or fails to leverage the generative ability of large-scale diffusion models. To address these issues, we introduce a novel artistic style transfer method based on a pre-trained large-scale diffusion model without any optimization. Specifically, we manipulate the features of self-attention layers as the way the cross-attention mechanism works; in the generation process, substituting the key and value of content with those of style image. This approach provides several desirable characteristics for style transfer including 1) preservation of content by transferring similar styles into similar image patches and 2) transfer of style based on similarity of local texture (e.g. edge) between content and style images. Furthermore, we introduce query preservation and attention temperature scaling to mitigate the issue of disruption of original content, and initial latent Adaptive Instance Normalization (AdaIN) to deal with the disharmonious color (failure to transfer the colors of style). Our experimental results demonstrate that our proposed method surpasses state-of-the-art methods in both conventional and diffusion-based style transfer baselines.
Authors: Yuan Sun, Xuan Wang, Yunfan Zhang, Jie Zhang, Caigui Jiang, Yu Guo, Fei Wang
Abstract: We present a method named iComMa to address the 6D camera pose estimation problem in computer vision. Conventional pose estimation methods typically rely on the target's CAD model or necessitate specific network training tailored to particular object classes. Some existing methods have achieved promising results in mesh-free object and scene pose estimation by inverting the Neural Radiance Fields (NeRF). However, they still struggle with adverse initializations such as large rotations and translations. To address this issue, we propose an efficient method for accurate camera pose estimation by inverting 3D Gaussian Splatting (3DGS). Specifically, a gradient-based differentiable framework optimizes camera pose by minimizing the residual between the query image and the rendered image, requiring no training. An end-to-end matching module is designed to enhance the model's robustness against adverse initializations, while minimizing pixel-level comparing loss aids in precise pose estimation. Experimental results on synthetic and complex real-world data demonstrate the effectiveness of the proposed approach in challenging conditions and the accuracy of camera pose estimation.
Authors: Jiayi Liu, Hou In Ivan Tam, Ali Mahdavi-Amiri, Manolis Savva
Abstract: We address the challenge of generating 3D articulated objects in a controllable fashion. Currently, modeling articulated 3D objects is either achieved through laborious manual authoring, or using methods from prior work that are hard to scale and control directly. We leverage the interplay between part shape, connectivity, and motion using a denoising diffusion-based method with attention modules designed to extract correlations between part attributes. Our method takes an object category label and a part connectivity graph as input and generates an object's geometry and motion parameters. The generated objects conform to user-specified constraints on the object category, part shape, and part articulation. Our experiments show that our method outperforms the state-of-the-art in articulated object generation, producing more realistic objects while conforming better to user constraints. Video Summary at: http://youtu.be/cH_rbKbyTpE
Authors: Yuming Gu, You Xie, Hongyi Xu, Guoxian Song, Yichun Shi, Di Chang, Jing Yang, Linjie Luo
Abstract: We present DiffPortrait3D, a conditional diffusion model that is capable of synthesizing 3D-consistent photo-realistic novel views from as few as a single in-the-wild portrait. Specifically, given a single RGB input, we aim to synthesize plausible but consistent facial details rendered from novel camera views with retained both identity and facial expression. In lieu of time-consuming optimization and fine-tuning, our zero-shot method generalizes well to arbitrary face portraits with unposed camera views, extreme facial expressions, and diverse artistic depictions. At its core, we leverage the generative prior of 2D diffusion models pre-trained on large-scale image datasets as our rendering backbone, while the denoising is guided with disentangled attentive control of appearance and camera pose. To achieve this, we first inject the appearance context from the reference image into the self-attention layers of the frozen UNets. The rendering view is then manipulated with a novel conditional control module that interprets the camera pose by watching a condition image of a crossed subject from the same view. Furthermore, we insert a trainable cross-view attention module to enhance view consistency, which is further strengthened with a novel 3D-aware noise generation process during inference. We demonstrate state-of-the-art results both qualitatively and quantitatively on our challenging in-the-wild and multi-view benchmarks.
Authors: Andong Lu, Jiacong Zhao, Chenglong Li, Jin Tang, Bin Luo
Abstract: Current RGBT tracking research relies on the complete multi-modal input, but modal information might miss due to some factors such as thermal sensor self-calibration and data transmission error, called modality-missing challenge in this work. To address this challenge, we propose a novel invertible prompt learning approach, which integrates the content-preserving prompts into a well-trained tracking model to adapt to various modality-missing scenarios, for robust RGBT tracking. Given one modality-missing scenario, we propose to utilize the available modality to generate the prompt of the missing modality to adapt to RGBT tracking model. However, the cross-modality gap between available and missing modalities usually causes semantic distortion and information loss in prompt generation. To handle this issue, we design the invertible prompter by incorporating the full reconstruction of the input available modality from the generated prompt. To provide a comprehensive evaluation platform, we construct several high-quality benchmark datasets, in which various modality-missing scenarios are considered to simulate real-world challenges. Extensive experiments on three modality-missing benchmark datasets show that our method achieves significant performance improvements compared with state-of-the-art methods. We have released the code and simulation datasets at: \href{https://github.com/Alexadlu/Modality-missing-RGBT-Tracking.git}{https://github.com/Alexadlu/Modality-missing-RGBT-Tracking.git}.
URLs: https://github.com/Alexadlu/Modality-missing-RGBT-Tracking.git, https://github.com/Alexadlu/Modality-missing-RGBT-Tracking.git
Authors: Yiran Song, Qianyu Zhou, Xiangtai Li, Deng-Ping Fan, Xuequan Lu, Lizhuang Ma
Abstract: In this paper, we address the challenge of image resolution variation for the Segment Anything Model (SAM). SAM, known for its zero-shot generalizability, exhibits a performance degradation when faced with datasets with varying image sizes. Previous approaches tend to resize the image to a fixed size or adopt structure modifications, hindering the preservation of SAM's rich prior knowledge. Besides, such task-specific tuning necessitates a complete retraining of the model, which is cost-expensive and unacceptable for deployment in the downstream tasks. In this paper, we reformulate this issue as a length extrapolation problem, where token sequence length varies while maintaining a consistent patch size for images of different sizes. To this end, we propose Scalable Bias-Mode Attention Mask (BA-SAM) to enhance SAM's adaptability to varying image resolutions while eliminating the need for structure modifications. Firstly, we introduce a new scaling factor to ensure consistent magnitude in the attention layer's dot product values when the token sequence length changes. Secondly, we present a bias-mode attention mask that allows each token to prioritize neighboring information, mitigating the impact of untrained distant information. Our BA-SAM demonstrates efficacy in two scenarios: zero-shot and fine-tuning. Extensive evaluation on diverse datasets, including DIS5K, DUTS, ISIC, COD10K, and COCO, reveals its ability to significantly mitigate performance degradation in the zero-shot setting and achieve state-of-the-art performance with minimal fine-tuning. Furthermore, we propose a generalized model and benchmark, showcasing BA-SAM's generalizability across all four datasets simultaneously. Code is available at https://github.com/zongzi13545329/BA-SAM
Authors: Xiangtao Kong, Chao Dong, Lei Zhang
Abstract: While single task image restoration (IR) has achieved significant successes, it remains a challenging issue to train a single model which can tackle multiple IR tasks. In this work, we investigate in-depth the multiple-in-one (MiO) IR problem, which comprises seven popular IR tasks. We point out that MiO IR faces two pivotal challenges: the optimization of diverse objectives and the adaptation to multiple tasks. To tackle these challenges, we present two simple yet effective strategies. The first strategy, referred to as sequential learning, attempts to address how to optimize the diverse objectives, which guides the network to incrementally learn individual IR tasks in a sequential manner rather than mixing them together. The second strategy, i.e., prompt learning, attempts to address how to adapt to the different IR tasks, which assists the network to understand the specific task and improves the generalization ability. By evaluating on 19 test sets, we demonstrate that the sequential and prompt learning strategies can significantly enhance the MiO performance of commonly used CNN and Transformer backbones. Our experiments also reveal that the two strategies can supplement each other to learn better degradation representations and enhance the model robustness. It is expected that our proposed MiO IR formulation and strategies could facilitate the research on how to train IR models with higher generalization capabilities.
Authors: Yuxuan Sun, Hao Wu, Chenglu Zhu, Sunyi Zheng, Qizi Chen, Kai Zhang, Yunlong Zhang, Dan Wan, Xiaoxiao Lan, Mengyue Zheng, Jingxiong Li, Xinheng Lyu, Tao Lin, Lin Yang
Abstract: The emergence of large multimodal models has unlocked remarkable potential in AI, particularly in pathology. However, the lack of specialized, high-quality benchmark impeded their development and precise evaluation. To address this, we introduce PathMMU, the largest and highest-quality expert-validated pathology benchmark for Large Multimodal Models (LMMs). It comprises 33,428 multimodal multi-choice questions and 24,067 images from various sources, each accompanied by an explanation for the correct answer. The construction of PathMMU harnesses GPT-4V's advanced capabilities, utilizing over 30,000 image-caption pairs to enrich captions and generate corresponding Q&As in a cascading process. Significantly, to maximize PathMMU's authority, we invite seven pathologists to scrutinize each question under strict standards in PathMMU's validation and test sets, while simultaneously setting an expert-level performance benchmark for PathMMU. We conduct extensive evaluations, including zero-shot assessments of 14 open-sourced and 4 closed-sourced LMMs and their robustness to image corruption. We also fine-tune representative LMMs to assess their adaptability to PathMMU. The empirical findings indicate that advanced LMMs struggle with the challenging PathMMU benchmark, with the top-performing LMM, GPT-4V, achieving only a 49.8% zero-shot performance, significantly lower than the 71.8% demonstrated by human pathologists. After fine-tuning, significantly smaller open-sourced LMMs can outperform GPT-4V but still fall short of the expertise shown by pathologists. We hope that the PathMMU will offer valuable insights and foster the development of more specialized, next-generation LMMs for pathology.
Authors: Shufan Li, Harkanwar Singh, Aditya Grover
Abstract: In recent years, Transformers have become the de-facto architecture for sequence modeling on text and a variety of multi-dimensional data, such as images and video. However, the use of self-attention layers in a Transformer incurs prohibitive compute and memory complexity that scales quadratically w.r.t. the sequence length. A recent architecture, Mamba, based on state space models has been shown to achieve comparable performance for modeling text sequences, while scaling linearly with the sequence length. In this work, we present Mamba-ND, a generalized design extending the Mamba architecture to arbitrary multi-dimensional data. Our design alternatively unravels the input data across different dimensions following row-major orderings. We provide a systematic comparison of Mamba-ND with several other alternatives, based on prior multi-dimensional extensions such as Bi-directional LSTMs and S4ND. Empirically, we show that Mamba-ND demonstrates performance competitive with the state-of-the-art on a variety of multi-dimensional benchmarks, including ImageNet-1K classification, HMDB-51 action recognition, and ERA5 weather forecasting.
Authors: Khang Truong Giang, Soohwan Song, Sungho Jo
Abstract: This study addresses the challenge of performing visual localization in demanding conditions such as night-time scenarios, adverse weather, and seasonal changes. While many prior studies have focused on improving image-matching performance to facilitate reliable dense keypoint matching between images, existing methods often heavily rely on predefined feature points on a reconstructed 3D model. Consequently, they tend to overlook unobserved keypoints during the matching process. Therefore, dense keypoint matches are not fully exploited, leading to a notable reduction in accuracy, particularly in noisy scenes. To tackle this issue, we propose a novel localization method that extracts reliable semi-dense 2D-3D matching points based on dense keypoint matches. This approach involves regressing semi-dense 2D keypoints into 3D scene coordinates using a point inference network. The network utilizes both geometric and visual cues to effectively infer 3D coordinates for unobserved keypoints from the observed ones. The abundance of matching information significantly enhances the accuracy of camera pose estimation, even in scenarios involving noisy or sparse 3D models. Comprehensive evaluations demonstrate that the proposed method outperforms other methods in challenging scenes and achieves competitive results in large-scale visual localization benchmarks. The code will be available.
Authors: Ze Ma, Daquan Zhou, Chun-Hsiao Yeh, Xue-She Wang, Xiuyu Li, Huanrui Yang, Zhen Dong, Kurt Keutzer, Jiashi Feng
Abstract: Creating content with specified identities (ID) has attracted significant interest in the field of generative models. In the field of text-to-image generation (T2I), subject-driven creation has achieved great progress with the identity controlled via reference images. However, its extension to video generation is not well explored. In this work, we propose a simple yet effective subject identity controllable video generation framework, termed Video Custom Diffusion (VCD). With a specified identity defined by a few images, VCD reinforces the identity characteristics and injects frame-wise correlation at the initialization stage for stable video outputs. To achieve this, we propose three novel components that are essential for high-quality identity preservation and stable video generation: 1) a noise initialization method with 3D Gaussian Noise Prior for better inter-frame stability; 2) an ID module based on extended Textual Inversion trained with the cropped identity to disentangle the ID information from the background 3) Face VCD and Tiled VCD modules to reinforce faces and upscale the video to higher resolution while preserving the identity's features. We conducted extensive experiments to verify that VCD is able to generate stable videos with better ID over the baselines. Besides, with the transferability of the encoded identity in the ID module, VCD is also working well with personalized text-to-image models available publicly. The codes are available at https://github.com/Zhen-Dong/Magic-Me.
Authors: Nguyen Nguyen, Jing Bi, Ali Vosoughi, Yapeng Tian, Pooyan Fazli, Chenliang Xu
Abstract: The capability of intelligent models to extrapolate and comprehend changes in object states is a crucial yet demanding aspect of AI research, particularly through the lens of human interaction in real-world settings. This task involves describing complex visual environments, identifying active objects, and interpreting their changes as conveyed through language. Traditional methods, which isolate object captioning and state change detection, offer a limited view of dynamic environments. Moreover, relying on a small set of symbolic words to represent changes has restricted the expressiveness of the language. To address these challenges, in this paper, we introduce the Object State Captioning and State Change Representation (OSCaR) dataset and benchmark. OSCaR consists of 14,084 annotated video segments with nearly 1,000 unique objects from various egocentric video collections. It sets a new testbed for evaluating multimodal large language models (MLLMs). Our experiments demonstrate that while MLLMs show some skill, they lack a full understanding of object state changes. The benchmark includes a fine-tuned model that, despite initial capabilities, requires significant improvements in accuracy and generalization ability for effective understanding of these changes. Our code and dataset are available at https://github.com/nguyennm1024/OSCaR.
Authors: Chanyoung Kim, Woojung Han, Dayun Ju, Seong Jae Hwang
Abstract: Semantic segmentation has innately relied on extensive pixel-level annotated data, leading to the emergence of unsupervised methodologies. Among them, leveraging self-supervised Vision Transformers for unsupervised semantic segmentation (USS) has been making steady progress with expressive deep features. Yet, for semantically segmenting images with complex objects, a predominant challenge remains: the lack of explicit object-level semantic encoding in patch-level features. This technical limitation often leads to inadequate segmentation of complex objects with diverse structures. To address this gap, we present a novel approach, EAGLE, which emphasizes object-centric representation learning for unsupervised semantic segmentation. Specifically, we introduce EiCue, a spectral technique providing semantic and structural cues through an eigenbasis derived from the semantic similarity matrix of deep image features and color affinity from an image. Further, by incorporating our object-centric contrastive loss with EiCue, we guide our model to learn object-level representations with intra- and inter-image object-feature consistency, thereby enhancing semantic accuracy. Extensive experiments on COCO-Stuff, Cityscapes, and Potsdam-3 datasets demonstrate the state-of-the-art USS results of EAGLE with accurate and consistent semantic segmentation across complex scenes.
Authors: Weiyi Lv, Yuhang Huang, Ning Zhang, Ruei-Sung Lin, Mei Han, Dan Zeng
Abstract: In Multiple Object Tracking, objects often exhibit non-linear motion of acceleration and deceleration, with irregular direction changes. Tacking-by-detection (TBD) trackers with Kalman Filter motion prediction work well in pedestrian-dominant scenarios but fall short in complex situations when multiple objects perform non-linear and diverse motion simultaneously. To tackle the complex non-linear motion, we propose a real-time diffusion-based MOT approach named DiffMOT. Specifically, for the motion predictor component, we propose a novel Decoupled Diffusion-based Motion Predictor (D$^2$MP). It models the entire distribution of various motion presented by the data as a whole. It also predicts an individual object's motion conditioning on an individual's historical motion information. Furthermore, it optimizes the diffusion process with much fewer sampling steps. As a MOT tracker, the DiffMOT is real-time at 22.7FPS, and also outperforms the state-of-the-art on DanceTrack and SportsMOT datasets with $62.3\%$ and $76.2\%$ in HOTA metrics, respectively. To the best of our knowledge, DiffMOT is the first to introduce a diffusion probabilistic model into the MOT to tackle non-linear motion prediction.
Authors: Imad Eddine Toubal, Aditya Avinash, Neil Gordon Alldrin, Jan Dlabal, Wenlei Zhou, Enming Luo, Otilia Stretcu, Hao Xiong, Chun-Ta Lu, Howard Zhou, Ranjay Krishna, Ariel Fuxman, Tom Duerig
Abstract: From content moderation to wildlife conservation, the number of applications that require models to recognize nuanced or subjective visual concepts is growing. Traditionally, developing classifiers for such concepts requires substantial manual effort measured in hours, days, or even months to identify and annotate data needed for training. Even with recently proposed Agile Modeling techniques, which enable rapid bootstrapping of image classifiers, users are still required to spend 30 minutes or more of monotonous, repetitive data labeling just to train a single classifier. Drawing on Fiske's Cognitive Miser theory, we propose a new framework that alleviates manual effort by replacing human labeling with natural language interactions, reducing the total effort required to define a concept by an order of magnitude: from labeling 2,000 images to only 100 plus some natural language interactions. Our framework leverages recent advances in foundation models, both large language models and vision-language models, to carve out the concept space through conversation and by automatically labeling training data points. Most importantly, our framework eliminates the need for crowd-sourced annotations. Moreover, our framework ultimately produces lightweight classification models that are deployable in cost-sensitive scenarios. Across 15 subjective concepts and across 2 public image classification datasets, our trained models outperform traditional Agile Modeling as well as state-of-the-art zero-shot classification models like ALIGN, CLIP, CuPL, and large visual question-answering models like PaLI-X.
Authors: Boeun Kim, Jungho Kim, Hyung Jin Chang, Jin Young Choi
Abstract: While existing motion style transfer methods are effective between two motions with identical content, their performance significantly diminishes when transferring style between motions with different contents. This challenge lies in the lack of clear separation between content and style of a motion. To tackle this challenge, we propose a novel motion style transformer that effectively disentangles style from content and generates a plausible motion with transferred style from a source motion. Our distinctive approach to achieving the goal of disentanglement is twofold: (1) a new architecture for motion style transformer with `part-attentive style modulator across body parts' and `Siamese encoders that encode style and content features separately'; (2) style disentanglement loss. Our method outperforms existing methods and demonstrates exceptionally high quality, particularly in motion pairs with different contents, without the need for heuristic post-processing. Codes are available at https://github.com/Boeun-Kim/MoST.
Authors: Xinhao Cai, Qiuxia Lai, Yuwei Wang, Wenguan Wang, Zeren Sun, Yazhou Yao
Abstract: Object detection in remote sensing images (RSIs) often suffers from several increasing challenges, including the large variation in object scales and the diverse-ranging context. Prior methods tried to address these challenges by expanding the spatial receptive field of the backbone, either through large-kernel convolution or dilated convolution. However, the former typically introduces considerable background noise, while the latter risks generating overly sparse feature representations. In this paper, we introduce the Poly Kernel Inception Network (PKINet) to handle the above challenges. PKINet employs multi-scale convolution kernels without dilation to extract object features of varying scales and capture local context. In addition, a Context Anchor Attention (CAA) module is introduced in parallel to capture long-range contextual information. These two components work jointly to advance the performance of PKINet on four challenging remote sensing detection benchmarks, namely DOTA-v1.0, DOTA-v1.5, HRSC2016, and DIOR-R.
Authors: Sergey Kastryulin, Denis Prokopenko, Artem Babenko, Dmitry V. Dylov
Abstract: This paper introduces a new data-driven, non-parametric method for image quality and aesthetics assessment, surpassing existing approaches and requiring no prompt engineering or fine-tuning. We eliminate the need for expressive textual embeddings by proposing efficient image anchors in the data. Through extensive evaluations of 7 state-of-the-art self-supervised models, our method demonstrates superior performance and robustness across various datasets and benchmarks. Notably, it achieves high agreement with human assessments even with limited data and shows high robustness to the nature of data and their pre-processing pipeline. Our contributions offer a streamlined solution for assessment of images while providing insights into the perception of visual information.
Authors: Pengze Zhang, Hubery Yin, Chen Li, Xiaohua Xie
Abstract: Most diffusion models assume that the reverse process adheres to a Gaussian distribution. However, this approximation has not been rigorously validated, especially at singularities, where t=0 and t=1. Improperly dealing with such singularities leads to an average brightness issue in applications, and limits the generation of images with extreme brightness or darkness. We primarily focus on tackling singularities from both theoretical and practical perspectives. Initially, we establish the error bounds for the reverse process approximation, and showcase its Gaussian characteristics at singularity time steps. Based on this theoretical insight, we confirm the singularity at t=1 is conditionally removable while it at t=0 is an inherent property. Upon these significant conclusions, we propose a novel plug-and-play method SingDiffusion to address the initial singular time step sampling, which not only effectively resolves the average brightness issue for a wide range of diffusion models without extra training efforts, but also enhances their generation capability in achieving notable lower FID scores.
Authors: Zihan Chu
Abstract: Adverse weather conditions including haze, snow and rain lead to decline in image qualities, which often causes a decline in performance for deep-learning based detection networks. Most existing approaches attempts to rectify hazy images before performing object detection, which increases the complexity of the network and may result in the loss in latent information. To better integrate image restoration and object detection tasks, we designed a double-route network with an attention feature fusion module, taking both hazy and dehazed features into consideration. We also proposed a subnetwork to provide haze-free features to the detection network. Specifically, our D-YOLO improves the performance of the detection network by minimizing the distance between the clear feature extraction subnetwork and detection network. Experiments on RTTS and FoggyCityscapes datasets show that D-YOLO demonstrates better performance compared to the state-of-the-art methods. It is a robust detection framework for bridging the gap between low-level dehazing and high-level detection.
Authors: Juan Tapia, Christoph Busch
Abstract: This paper evaluated the impact of synthetic images on Morphing Attack Detection (MAD) using a Siamese network with a semi-hard-loss function. Intra and cross-dataset evaluations were performed to measure synthetic image generalisation capabilities using a cross-dataset for evaluation. Three different pre-trained networks were used as feature extractors from traditional MobileNetV2, MobileNetV3 and EfficientNetB0. Our results show that MAD trained on EfficientNetB0 from FERET, FRGCv2, and FRLL can reach a lower error rate in comparison with SOTA. Conversely, worse performances were reached when the system was trained only with synthetic images. A mixed approach (synthetic + digital) database may help to improve MAD and reduce the error rate. This fact shows that we still need to keep going with our efforts to include synthetic images in the training process.
Authors: Ruida Zhang, Chenyangguang Zhang, Yan Di, Fabian Manhardt, Xingyu Liu, Federico Tombari, Xiangyang Ji
Abstract: In this paper, we present KP-RED, a unified KeyPoint-driven REtrieval and Deformation framework that takes object scans as input and jointly retrieves and deforms the most geometrically similar CAD models from a pre-processed database to tightly match the target. Unlike existing dense matching based methods that typically struggle with noisy partial scans, we propose to leverage category-consistent sparse keypoints to naturally handle both full and partial object scans. Specifically, we first employ a lightweight retrieval module to establish a keypoint-based embedding space, measuring the similarity among objects by dynamically aggregating deformation-aware local-global features around extracted keypoints. Objects that are close in the embedding space are considered similar in geometry. Then we introduce the neural cage-based deformation module that estimates the influence vector of each keypoint upon cage vertices inside its local support region to control the deformation of the retrieved shape. Extensive experiments on the synthetic dataset PartNet and the real-world dataset Scan2CAD demonstrate that KP-RED surpasses existing state-of-the-art approaches by a large margin. Codes and trained models will be released in https://github.com/lolrudy/KP-RED.
Authors: Yuting Xu, Jian Liang, Lijun Sheng, Xiao-Yu Zhang
Abstract: The deepfake threats to society and cybersecurity have provoked significant public apprehension, driving intensified efforts within the realm of deepfake video detection. Current video-level methods are mostly based on {3D CNNs} resulting in high computational demands, although have achieved good performance. This paper introduces an elegantly simple yet effective strategy named Thumbnail Layout (TALL), which transforms a video clip into a pre-defined layout to realize the preservation of spatial and temporal dependencies. This transformation process involves sequentially masking frames at the same positions within each frame. These frames are then resized into sub-frames and reorganized into the predetermined layout, forming thumbnails. TALL is model-agnostic and has remarkable simplicity, necessitating only minimal code modifications. Furthermore, we introduce a graph reasoning block (GRB) and semantic consistency (SC) loss to strengthen TALL, culminating in TALL++. GRB enhances interactions between different semantic regions to capture semantic-level inconsistency clues. The semantic consistency loss imposes consistency constraints on semantic features to improve model generalization ability. Extensive experiments on intra-dataset, cross-dataset, diffusion-generated image detection, and deepfake generation method recognition show that TALL++ achieves results surpassing or comparable to the state-of-the-art methods, demonstrating the effectiveness of our approaches for various deepfake detection problems. The code is available at https://github.com/rainy-xu/TALL4Deepfake.
Authors: Huy-Hoang Bui, Bach-Thuan Bui, Dinh-Tuan Tran, Joo-Ho Lee
Abstract: Classical structural-based visual localization methods offer high accuracy but face trade-offs in terms of storage, speed, and privacy. A recent innovation, keypoint scene coordinate regression (KSCR) named D2S addresses these issues by leveraging graph attention networks to enhance keypoint relationships and predict their 3D coordinates using a simple multilayer perceptron (MLP). Camera pose is then determined via PnP+RANSAC, using established 2D-3D correspondences. While KSCR achieves competitive results, rivaling state-of-the-art image-retrieval methods like HLoc across multiple benchmarks, its performance is hindered when data samples are limited due to the deep learning model's reliance on extensive data. This paper proposes a solution to this challenge by introducing a pipeline for keypoint descriptor synthesis using Neural Radiance Field (NeRF). By generating novel poses and feeding them into a trained NeRF model to create new views, our approach enhances the KSCR's generalization capabilities in data-scarce environments. The proposed system could significantly improve localization accuracy by up to 50% and cost only a fraction of time for data synthesis. Furthermore, its modular design allows for the integration of multiple NeRFs, offering a versatile and efficient solution for visual localization. The implementation is publicly available at: https://github.com/ais-lab/DescriptorSynthesis4Feat2Map.
URLs: https://github.com/ais-lab/DescriptorSynthesis4Feat2Map.
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 the auto manufacturing and broader machine learning and computer vision communities using OSPD to advance the state of the art in defect detection of stamped holes in the metalsheet stamping process. The dataset is available for download at: https://tinyurl.com/hm6xatd7
Authors: Zixian Ma, Weikai Huang, Jieyu Zhang, Tanmay Gupta, Ranjay Krishna
Abstract: Real-world multi-modal problems are rarely solved by a single machine learning model, and often require multi-step computational plans that involve stitching several models. Tool-augmented LLMs hold tremendous promise for automating the generation of such computational plans. However, the lack of standardized benchmarks for evaluating LLMs as planners for multi-step multi-modal tasks has prevented a systematic study of planner design decisions. Should LLMs generate a full plan in a single shot or step-by-step? Should they invoke tools directly with Python code or through structured data formats like JSON? Does feedback improve planning? To answer these questions and more, we introduce m&m's: a benchmark containing 4K+ multi-step multi-modal tasks involving 33 tools that include multi-modal models, (free) public APIs, and image processing modules. For each of these task queries, we provide automatically generated plans using this realistic toolset. We further provide a high-quality subset of 1,565 task plans that are human-verified and correctly executable. With m&m's, we evaluate 6 popular LLMs with 2 planning strategies (multi-step vs. step-by-step planning), 2 plan formats (JSON vs. code), and 3 types of feedback (parsing/verification/execution). Finally, we summarize takeaways from our extensive experiments. Our dataset and code are available on HuggingFace (https://huggingface.co/datasets/zixianma/mnms) and Github (https://github.com/RAIVNLab/mnms).
URLs: https://huggingface.co/datasets/zixianma/mnms), https://github.com/RAIVNLab/mnms).
Authors: Qucheng Peng, Ce Zheng, Chen Chen
Abstract: 3D human pose data collected in controlled laboratory settings present challenges for pose estimators that generalize across diverse scenarios. To address this, domain generalization is employed. Current methodologies in domain generalization for 3D human pose estimation typically utilize adversarial training to generate synthetic poses for training. Nonetheless, these approaches exhibit several limitations. First, the lack of prior information about the target domain complicates the application of suitable augmentation through a single pose augmentor, affecting generalization on target domains. Moreover, adversarial training's discriminator tends to enforce similarity between source and synthesized poses, impeding the exploration of out-of-source distributions. Furthermore, the pose estimator's optimization is not exposed to domain shifts, limiting its overall generalization ability. To address these limitations, we propose a novel framework featuring two pose augmentors: the weak and the strong augmentors. Our framework employs differential strategies for generation and discrimination processes, facilitating the preservation of knowledge related to source poses and the exploration of out-of-source distributions without prior information about target poses. Besides, we leverage meta-optimization to simulate domain shifts in the optimization process of the pose estimator, thereby improving its generalization ability. Our proposed approach significantly outperforms existing methods, as demonstrated through comprehensive experiments on various benchmark datasets.Our code will be released at \url{https://github.com/davidpengucf/DAF-DG}.
Authors: Baolu Li, Jinlong Li, Xinyu Liu, Runsheng Xu, Zhengzhong Tu, Jiacheng Guo, Xiaopeng Li, Hongkai Yu
Abstract: Current LiDAR-based Vehicle-to-Everything (V2X) multi-agent perception systems have shown the significant success on 3D object detection. While these models perform well in the trained clean weather, they struggle in unseen adverse weather conditions with the real-world domain gap. In this paper, we propose a domain generalization approach, named V2X-DGW, for LiDAR-based 3D object detection on multi-agent perception system under adverse weather conditions. Not only in the clean weather does our research aim to ensure favorable multi-agent performance, but also in the unseen adverse weather conditions by learning only on the clean weather data. To advance research in this area, we have simulated the impact of three prevalent adverse weather conditions on two widely-used multi-agent datasets, resulting in the creation of two novel benchmark datasets: OPV2V-w and V2XSet-w. To this end, we first introduce the Adaptive Weather Augmentation (AWA) to mimic the unseen adverse weather conditions, and then propose two alignments for generalizable representation learning: Trust-region Weather-invariant Alignment (TWA) and Agent-aware Contrastive Alignment (ACA). Extensive experimental results demonstrate that our V2X-DGW achieved improvements in the unseen adverse weather conditions.
Authors: Mincheol Chang, Siyeong Lee, Jinkyu Kim, Namil Kim
Abstract: Typical LiDAR-based 3D object detection models are trained in a supervised manner with real-world data collection, which is often imbalanced over classes (or long-tailed). To deal with it, augmenting minority-class examples by sampling ground truth (GT) LiDAR points from a database and pasting them into a scene of interest is often used, but challenges still remain: inflexibility in locating GT samples and limited sample diversity. In this work, we propose to leverage pseudo-LiDAR point clouds generated (at a low cost) from videos capturing a surround view of miniatures or real-world objects of minor classes. Our method, called Pseudo Ground Truth Augmentation (PGT-Aug), consists of three main steps: (i) volumetric 3D instance reconstruction using a 2D-to-3D view synthesis model, (ii) object-level domain alignment with LiDAR intensity estimation and (iii) a hybrid context-aware placement method from ground and map information. We demonstrate the superiority and generality of our method through performance improvements in extensive experiments conducted on three popular benchmarks, i.e., nuScenes, KITTI, and Lyft, especially for the datasets with large domain gaps captured by different LiDAR configurations. Our code and data will be publicly available upon publication.
Authors: Xiaohan Lei, Min Wang, Wengang Zhou, Houqiang Li
Abstract: In embodied vision, Instance ImageGoal Navigation (IIN) requires an agent to locate a specific object depicted in a goal image within an unexplored environment. The primary difficulty of IIN stems from the necessity of recognizing the target object across varying viewpoints and rejecting potential distractors. Existing map-based navigation methods largely adopt the representation form of Bird's Eye View (BEV) maps, which, however, lack the representation of detailed textures in a scene. To address the above issues, we propose a new Gaussian Splatting Navigation (abbreviated as GaussNav) framework for IIN task, which constructs a novel map representation based on 3D Gaussian Splatting (3DGS). The proposed framework enables the agent to not only memorize the geometry and semantic information of the scene, but also retain the textural features of objects. Our GaussNav framework demonstrates a significant leap in performance, evidenced by an increase in Success weighted by Path Length (SPL) from 0.252 to 0.578 on the challenging Habitat-Matterport 3D (HM3D) dataset. Our code will be made publicly available.
Authors: Hang Wang, Zhi-Qi Cheng, Youtian Du, Lei Zhang
Abstract: Video Action Counting (VAC) is crucial in analyzing sports, fitness, and everyday activities by quantifying repetitive actions in videos. However, traditional VAC methods have overlooked the complexity of action repetitions, such as interruptions and the variability in cycle duration. Our research addresses the shortfall by introducing a novel approach to VAC, called Irregular Video Action Counting (IVAC). IVAC prioritizes modeling irregular repetition patterns in videos, which we define through two primary aspects: Inter-cycle Consistency and Cycle-interval Inconsistency. Inter-cycle Consistency ensures homogeneity in the spatial-temporal representations of cycle segments, signifying action uniformity within cycles. Cycle-interval inconsistency highlights the importance of distinguishing between cycle segments and intervals based on their inherent content differences. To encapsulate these principles, we propose a new methodology that includes consistency and inconsistency modules, supported by a unique pull-push loss (P2L) mechanism. The IVAC-P2L model applies a pull loss to promote coherence among cycle segment features and a push loss to clearly distinguish features of cycle segments from interval segments. Empirical evaluations conducted on the RepCount dataset demonstrate that the IVAC-P2L model sets a new benchmark in VAC task performance. Furthermore, the model demonstrates exceptional adaptability and generalization across various video contents, outperforming existing models on two additional datasets, UCFRep and Countix, without the need for dataset-specific optimization. These results confirm the efficacy of our approach in addressing irregular repetitions in videos and pave the way for further advancements in video analysis and understanding.
Authors: Rui Ma, Qiang Zhou, Bangjun Xiao, Yizhu Jin, Daquan Zhou, Xiuyu Li, Aishani Singh, Yi Qu, Kurt Keutzer, Xiaodong Xie, Jingtong Hu, Zhen Dong, Shanghang Zhang
Abstract: Copyright is a legal right that grants creators the exclusive authority to reproduce, distribute, and profit from their creative works. However, the recent advancements in text-to-image generation techniques have posed significant challenges to copyright protection, as these methods have facilitated the learning of unauthorized content, artistic creations, and portraits, which are subsequently utilized to generate and disseminate uncontrolled content. Especially, the use of stable diffusion, an emerging model for text-to-image generation, poses an increased risk of unauthorized copyright infringement and distribution. Currently, there is a lack of systematic studies evaluating the potential correlation between content generated by stable diffusion and those under copyright protection. Conducting such studies faces several challenges, including i) the intrinsic ambiguity related to copyright infringement in text-to-image models, ii) the absence of a comprehensive large-scale dataset, and iii) the lack of standardized metrics for defining copyright infringement. This work provides the first large-scale standardized dataset and benchmark on copyright protection. Specifically, we propose a pipeline to coordinate CLIP, ChatGPT, and diffusion models to generate a dataset that contains anchor images, corresponding prompts, and images generated by text-to-image models, reflecting the potential abuses of copyright. Furthermore, we explore a suite of evaluation metrics to judge the effectiveness of copyright protection methods. The proposed dataset, benchmark library, and evaluation metrics will be open-sourced to facilitate future research and application. The website and dataset can be accessed website dataset.
Authors: Siddharth Joshi, Arnav Jain, Ali Payani, Baharan Mirzasoleiman
Abstract: Contrastive Language-Image Pre-training (CLIP) on large-scale image-caption datasets learns representations that can achieve remarkable zero-shot generalization. However, such models require a massive amount of pre-training data. Improving the quality of the pre-training data has been shown to be much more effective in improving CLIP's performance than increasing its volume. Nevertheless, finding small subsets of training data that provably generalize the best has remained an open question. In this work, we propose the first theoretically rigorous data selection method for CLIP. We show that subsets that closely preserve the cross-covariance of the images and captions of the full data provably achieve a superior generalization performance. Our extensive experiments on ConceptualCaptions3M and ConceptualCaptions12M demonstrate that subsets found by \method\ achieve over 2.7x and 1.4x the accuracy of the next best baseline on ImageNet and its shifted versions. Moreover, we show that our subsets obtain 1.5x the average accuracy across 11 downstream datasets, of the next best baseline. The code is available at: https://github.com/BigML-CS-UCLA/clipcov-data-efficient-clip.
URLs: https://github.com/BigML-CS-UCLA/clipcov-data-efficient-clip.
Authors: Jun Yu, Gongpeng Zhao, Yongqi Wang, Zhihong Wei, Yang Zheng, Zerui Zhang, Zhongpeng Cai, Guochen Xie, Jichao Zhu, Wangyuan Zhu
Abstract: This paper presents our approach for the VA (Valence-Arousal) estimation task in the ABAW6 competition. We devised a comprehensive model by preprocessing video frames and audio segments to extract visual and audio features. Through the utilization of Temporal Convolutional Network (TCN) modules, we effectively captured the temporal and spatial correlations between these features. Subsequently, we employed a Transformer encoder structure to learn long-range dependencies, thereby enhancing the model's performance and generalization ability. Our method leverages a multimodal data fusion approach, integrating pre-trained audio and video backbones for feature extraction, followed by TCN-based spatiotemporal encoding and Transformer-based temporal information capture. Experimental results demonstrate the effectiveness of our approach, achieving competitive performance in VA estimation on the AffWild2 dataset.
Authors: Xiaoben Li, Mancheng Meng, Ziyan Wu, Terrence Chen, Fan Yang, Dinggang Shen
Abstract: Human mesh recovery from arbitrary multi-view images involves two characteristics: the arbitrary camera poses and arbitrary number of camera views. Because of the variability, designing a unified framework to tackle this task is challenging. The challenges can be summarized as the dilemma of being able to simultaneously estimate arbitrary camera poses and recover human mesh from arbitrary multi-view images while maintaining flexibility. To solve this dilemma, we propose a divide and conquer framework for Unified Human Mesh Recovery (U-HMR) from arbitrary multi-view images. In particular, U-HMR consists of a decoupled structure and two main components: camera and body decoupling (CBD), camera pose estimation (CPE), and arbitrary view fusion (AVF). As camera poses and human body mesh are independent of each other, CBD splits the estimation of them into two sub-tasks for two individual sub-networks (\ie, CPE and AVF) to handle respectively, thus the two sub-tasks are disentangled. In CPE, since each camera pose is unrelated to the others, we adopt a shared MLP to process all views in a parallel way. In AVF, in order to fuse multi-view information and make the fusion operation independent of the number of views, we introduce a transformer decoder with a SMPL parameters query token to extract cross-view features for mesh recovery. To demonstrate the efficacy and flexibility of the proposed framework and effect of each component, we conduct extensive experiments on three public datasets: Human3.6M, MPI-INF-3DHP, and TotalCapture.
Authors: Aakash Singh, Vivek Kumar Singh
Abstract: The advancements in computer vision and image processing techniques have led to emergence of new application in the domain of visual surveillance, targeted advertisement, content-based searching, and human-computer interaction etc. Out of the various techniques in computer vision, face analysis, in particular, has gained much attention. Several previous studies have tried to explore different applications of facial feature processing for a variety of tasks, including age and gender classification. However, despite several previous studies having explored the problem, the age and gender classification of in-wild human faces is still far from the achieving the desired levels of accuracy required for real-world applications. This paper, therefore, attempts to bridge this gap by proposing a hybrid model that combines self-attention and BiLSTM approaches for age and gender classification problems. The proposed models performance is compared with several state-of-the-art model proposed so far. An improvement of approximately 10percent and 6percent over the state-of-the-art implementations for age and gender classification, respectively, are noted for the proposed model. The proposed model is thus found to achieve superior performance and is found to provide a more generalized learning. The model can, therefore, be applied as a core classification component in various image processing and computer vision problems.
Authors: Yuan Gao, Sangwook Kim, David E Austin, Chris McIntosh
Abstract: Medical vision-language pretraining models (VLPM) have achieved remarkable progress in fusing chest X-rays (CXR) with clinical texts, introducing image-text data binding approaches that enable zero-shot learning and downstream clinical tasks. However, the current landscape lacks the holistic integration of additional medical modalities, such as electrocardiograms (ECG). We present MEDBind (Medical Electronic patient recorD), which learns joint embeddings across CXR, ECG, and medical text. Using text data as the central anchor, MEDBind features tri-modality binding, delivering competitive performance in top-K retrieval, zero-shot, and few-shot benchmarks against established VLPM, and the ability for CXR-to-ECG zero-shot classification and retrieval. This seamless integration is achieved through combination of contrastive loss on modality-text pairs with our proposed contrastive loss function, Edge-Modality Contrastive Loss, fostering a cohesive embedding space for CXR, ECG, and text. Finally, we demonstrate that MEDBind can improve downstream tasks by directly integrating CXR and ECG embeddings into a large-language model for multimodal prompt tuning.
Authors: Mihir Parmar, Swaroop Mishra, Mor Geva, Chitta Baral
Abstract: In recent years, progress in NLU has been driven by benchmarks. These benchmarks are typically collected by crowdsourcing, where annotators write examples based on annotation instructions crafted by dataset creators. In this work, we hypothesize that annotators pick up on patterns in the crowdsourcing instructions, which bias them to write many similar examples that are then over-represented in the collected data. We study this form of bias, termed instruction bias, in 14 recent NLU benchmarks, showing that instruction examples often exhibit concrete patterns, which are propagated by crowdworkers to the collected data. This extends previous work (Geva et al., 2019) and raises a new concern of whether we are modeling the dataset creator's instructions, rather than the task. Through a series of experiments, we show that, indeed, instruction bias can lead to overestimation of model performance, and that models struggle to generalize beyond biases originating in the crowdsourcing instructions. We further analyze the influence of instruction bias in terms of pattern frequency and model size, and derive concrete recommendations for creating future NLU benchmarks.
Authors: Xiao Li, Qiongxiu Li, Zhanhao Hu, Xiaolin Hu
Abstract: Machine learning poses severe privacy concerns as it has been shown that the learned models can reveal sensitive information about their training data. Many works have investigated the effect of widely adopted data augmentation and adversarial training techniques, termed data enhancement in the paper, on the privacy leakage of machine learning models. Such privacy effects are often measured by membership inference attacks (MIAs), which aim to identify whether a particular example belongs to the training set or not. We propose to investigate privacy from a new perspective called memorization. Through the lens of memorization, we find that previously deployed MIAs produce misleading results as they are less likely to identify samples with higher privacy risks as members compared to samples with low privacy risks. To solve this problem, we deploy a recent attack that can capture individual samples' memorization degrees for evaluation. Through extensive experiments, we unveil several findings about the connections between three essential properties of machine learning models, including privacy, generalization gap, and adversarial robustness. We demonstrate that the generalization gap and privacy leakage are less correlated than those of the previous results. Moreover, there is not necessarily a trade-off between adversarial robustness and privacy as stronger adversarial robustness does not make the model more susceptible to privacy attacks.
Authors: Saddam Hussain Khan (Department of Computer Systems Engineering, University of Engineering and Applied Science, Swat, Pakistan), Tahani Jaser Alahmadi (Department of Information Systems, Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia)
Abstract: Malaria is a potentially fatal plasmodium parasite injected by female anopheles mosquitoes that infect red blood cells and millions worldwide yearly. However, specialists' manual screening in clinical practice is laborious and prone to error. Therefore, a novel Deep Boosted and Ensemble Learning (DBEL) framework, comprising the stacking of new Boosted-BR-STM convolutional neural networks (CNN) and the ensemble ML classifiers, is developed to screen malaria parasite images. The proposed Boosted-BR-STM is based on a new dilated-convolutional block-based split transform merge (STM) and feature-map Squeezing-Boosting (SB) ideas. Moreover, the new STM block uses regional and boundary operations to learn the malaria parasite's homogeneity, heterogeneity, and boundary with patterns. Furthermore, the diverse boosted channels are attained by employing Transfer Learning-based new feature-map SB in STM blocks at the abstract, medium, and conclusion levels to learn minute intensity and texture variation of the parasitic pattern. The proposed DBEL framework implicates the stacking of prominent and diverse boosted channels and provides the generated discriminative features of the developed Boosted-BR-STM to the ensemble of ML classifiers. The proposed framework improves the discrimination ability and generalization of ensemble learning. Moreover, the deep feature spaces of the developed Boosted-BR-STM and customized CNNs are fed into ML classifiers for comparative analysis. The proposed DBEL framework outperforms the existing techniques on the NIH malaria dataset that are enhanced using discrete wavelet transform to enrich feature space. The proposed DBEL framework achieved Accuracy (98.50%), Sensitivity (0.9920), F-score (0.9850), and AUC (0.997), which suggest it to be utilized for malaria parasite screening.
Authors: Jaskaran Singh Walia, Aryan Odugoudar
Abstract: Several websites improve their security and avoid dangerous Internet attacks by implementing CAPTCHAs (Completely Automated Public Turing test to tell Computers and Humans Apart), a type of verification to identify whether the end-user is human or a robot. The most prevalent type of CAPTCHA is text-based, designed to be easily recognized by humans while being unsolvable towards machines or robots. However, as deep learning technology progresses, development of convolutional neural network (CNN) models that predict text-based CAPTCHAs becomes easier. The purpose of this research is to investigate the flaws and vulnerabilities in the CAPTCHA generating systems in order to design more resilient CAPTCHAs. To achieve this, we created CapNet, a Convolutional Neural Network. The proposed platform can evaluate both numerical and alphanumerical CAPTCHAs
Authors: Ashish Sinha, Jeremy Kawahara, Arezou Pakzad, Kumar Abhishek, Matthieu Ruthven, Enjie Ghorbel, Anis Kacem, Djamila Aouada, Ghassan Hamarneh
Abstract: In recent years, deep learning (DL) has shown great potential in the field of dermatological image analysis. However, existing datasets in this domain have significant limitations, including a small number of image samples, limited disease conditions, insufficient annotations, and non-standardized image acquisitions. To address these shortcomings, we propose a novel framework called DermSynth3D. DermSynth3D blends skin disease patterns onto 3D textured meshes of human subjects using a differentiable renderer and generates 2D images from various camera viewpoints under chosen lighting conditions in diverse background scenes. Our method adheres to top-down rules that constrain the blending and rendering process to create 2D images with skin conditions that mimic in-the-wild acquisitions, ensuring more meaningful results. The framework generates photo-realistic 2D dermoscopy images and the corresponding dense annotations for semantic segmentation of the skin, skin conditions, body parts, bounding boxes around lesions, depth maps, and other 3D scene parameters, such as camera position and lighting conditions. DermSynth3D allows for the creation of custom datasets for various dermatology tasks. We demonstrate the effectiveness of data generated using DermSynth3D by training DL models on synthetic data and evaluating them on various dermatology tasks using real 2D dermatological images. We make our code publicly available at https://github.com/sfu-mial/DermSynth3D.
Authors: Pengzhen Ren, Kaidong Zhang, Hetao Zheng, Zixuan Li, Yuhang Wen, Fengda Zhu, Mas Ma, Xiaodan Liang
Abstract: Considering how to make the model accurately understand and follow natural language instructions and perform actions consistent with world knowledge is a key challenge in robot manipulation. This mainly includes human fuzzy instruction reasoning and the following of physical knowledge. Therefore, the embodied intelligence agent must have the ability to model world knowledge from training data. However, most existing vision and language robot manipulation methods mainly operate in less realistic simulator and language settings and lack explicit modeling of world knowledge. To bridge this gap, we introduce a novel and simple robot manipulation framework, called Surfer. It is based on the world model, treats robot manipulation as a state transfer of the visual scene, and decouples it into two parts: action and scene. Then, the generalization ability of the model on new instructions and new scenes is enhanced by explicit modeling of the action and scene prediction in multi-modal information. In addition to the framework, we also built a robot manipulation simulator that supports full physics execution based on the MuJoCo physics engine. It can automatically generate demonstration training data and test data, effectively reducing labor costs. To conduct a comprehensive and systematic evaluation of the robot manipulation model in terms of language understanding and physical execution, we also created a robotic manipulation benchmark with progressive reasoning tasks, called SeaWave. It contains 4 levels of progressive reasoning tasks and can provide a standardized testing platform for embedded AI agents in multi-modal environments. On average, Surfer achieved a success rate of 54.74% on the defined four levels of manipulation tasks, exceeding the best baseline performance of 47.64%.
Authors: Amirhossein Rasoulian, Arash Harirpoush, Soorena Salari, Yiming Xiao
Abstract: Accurate identification and quantification of unruptured intracranial aneurysms (UIAs) is crucial for the risk assessment and treatment of this cerebrovascular disorder. Current 2D manual assessment on 3D magnetic resonance angiography (MRA) is suboptimal and time-consuming. In addition, one major issue in medical image segmentation is the need for large well-annotated data, which can be expensive to obtain. Techniques that mitigate this requirement, such as weakly supervised learning with coarse labels are highly desirable. In the paper, we propose FocalSegNet, a novel 3D focal modulation UNet, to detect an aneurysm and offer an initial, coarse segmentation of it from time-of-flight MRA image patches, which is further refined with a dense conditional random field (CRF) post-processing layer to produce a final segmentation map. We trained and evaluated our model on a public dataset, and in terms of UIA detection, our model showed a low false-positive rate of 0.21 and a high sensitivity of 0.80. For voxel-wise aneurysm segmentation, we achieved a Dice score of 0.68 and a 95% Hausdorff distance of ~0.95 mm, demonstrating its strong performance. We evaluated our algorithms against the state-of-the-art 3D Residual-UNet and Swin-UNETR, and illustrated the superior performance of our proposed FocalSegNet, highlighting the advantages of employing focal modulation for this task.
Authors: Yunjie Chen, Marius Staring, Olaf M. Neve, Stephan R. Romeijn, Erik F. Hensen, Berit M. Verbist, Jelmer M. Wolterink, Qian Tao
Abstract: Multi-sequence magnetic resonance imaging (MRI) has found wide applications in both modern clinical studies and deep learning research. However, in clinical practice, it frequently occurs that one or more of the MRI sequences are missing due to different image acquisition protocols or contrast agent contraindications of patients, limiting the utilization of deep learning models trained on multi-sequence data. One promising approach is to leverage generative models to synthesize the missing sequences, which can serve as a surrogate acquisition. State-of-the-art methods tackling this problem are based on convolutional neural networks (CNN) which usually suffer from spectral biases, resulting in poor reconstruction of high-frequency fine details. In this paper, we propose Conditional Neural fields with Shift modulation (CoNeS), a model that takes voxel coordinates as input and learns a representation of the target images for multi-sequence MRI translation. The proposed model uses a multi-layer perceptron (MLP) instead of a CNN as the decoder for pixel-to-pixel mapping. Hence, each target image is represented as a neural field that is conditioned on the source image via shift modulation with a learned latent code. Experiments on BraTS 2018 and an in-house clinical dataset of vestibular schwannoma patients showed that the proposed method outperformed state-of-the-art methods for multi-sequence MRI translation both visually and quantitatively. Moreover, we conducted spectral analysis, showing that CoNeS was able to overcome the spectral bias issue common in conventional CNN models. To further evaluate the usage of synthesized images in clinical downstream tasks, we tested a segmentation network using the synthesized images at inference.
Authors: Yangyi Chen, Karan Sikka, Michael Cogswell, Heng Ji, Ajay Divakaran
Abstract: Vision-language models (VLMs) have recently demonstrated strong efficacy as visual assistants that can parse natural queries about the visual content and generate human-like outputs. In this work, we explore the ability of these models to demonstrate human-like reasoning based on the perceived information. To address a crucial concern regarding the extent to which their reasoning capabilities are fully consistent and grounded, we also measure the reasoning consistency of these models. We achieve this by proposing a chain-of-thought (CoT) based consistency measure. However, such an evaluation requires a benchmark that encompasses both high-level inference and detailed reasoning chains, which is costly. We tackle this challenge by proposing a LLM-Human-in-the-Loop pipeline, which notably reduces cost while simultaneously ensuring the generation of a high-quality dataset. Based on this pipeline and the existing coarse-grained annotated dataset, we build the CURE benchmark to measure both the zero-shot reasoning performance and consistency of VLMs. We evaluate existing state-of-the-art VLMs, and find that even the best-performing model is unable to demonstrate strong visual reasoning capabilities and consistency, indicating that substantial efforts are required to enable VLMs to perform visual reasoning as systematically and consistently as humans. As an early step, we propose a two-stage training framework aimed at improving both the reasoning performance and consistency of VLMs. The first stage involves employing supervised fine-tuning of VLMs using step-by-step reasoning samples automatically generated by LLMs. In the second stage, we further augment the training process by incorporating feedback provided by LLMs to produce reasoning chains that are highly consistent and grounded. We empirically highlight the effectiveness of our framework in both reasoning performance and consistency.
Authors: Haozhe Zhao, Zefan Cai, Shuzheng Si, Xiaojian Ma, Kaikai An, Liang Chen, Zixuan Liu, Sheng Wang, Wenjuan Han, Baobao Chang
Abstract: Since the resurgence of deep learning, vision-language models (VLMs) enhanced by large language models (LLMs) have grown exponentially in popularity. However, while LLMs can utilize extensive background knowledge and task information with in-context learning, most VLMs still struggle with understanding complex multi-modal prompts with multiple images, making VLMs less effective in downstream vision-language tasks. In this paper, we address the limitation above by 1) introducing vision-language Model with Multi-Modal In-Context Learning(MMICL), a new approach to allow the VLM to deal with multi-modal inputs efficiently; 2) proposing a novel context scheme to augment the in-context learning ability of the VLM; 3) constructing the Multi-modal In-Context Learning (MIC) dataset, designed to enhance the VLM's ability to understand complex multi-modal prompts. Our experiments confirm that MMICL achieves new state-of-the-art zero-shot performance on a wide range of general vision-language tasks, especially for complex benchmarks, including MME and MMBench. Our analysis demonstrates that MMICL effectively tackles the challenge of complex multi-modal prompt understanding and emerges the impressive ICL ability. Furthermore, we observe that MMICL successfully alleviates language bias in VLMs, a common issue for VLMs that often leads to hallucination when faced with extensive textual context. Our code, dataset, dataset tool, and model are available at https://github.com/PKUnlp-icler/MIC
Authors: Xuechun Li, Paula M. Burgi, Wei Ma, Hae Young Noh, David J. Wald, Susu Xu
Abstract: Onsite disasters like earthquakes can trigger cascading hazards and impacts, such as landslides and infrastructure damage, leading to catastrophic losses; thus, rapid and accurate estimates are crucial for timely and effective post-disaster responses. Interferometric Synthetic aperture radar (InSAR) data is important in providing high-resolution onsite information for rapid hazard estimation. Most recent methods using InSAR imagery signals predict a single type of hazard and thus often suffer low accuracy due to noisy and complex signals induced by co-located hazards, impacts, and irrelevant environmental changes (e.g., vegetation changes, human activities). We introduce a novel stochastic variational inference with normalizing flows derived to jointly approximate posteriors of multiple unobserved hazards and impacts from noisy InSAR imagery.
Authors: Maren H{\o}ib{\o}, Andr\'e Pedersen, Vibeke Grotnes Dale, Sissel Marie Berget, Borgny Ytterhus, Cecilia Lindskog, Elisabeth Wik, Lars A. Akslen, Ingerid Reinertsen, Erik Smistad, Marit Valla
Abstract: Digital pathology enables automatic analysis of histopathological sections using artificial intelligence (AI). Automatic evaluation could improve diagnostic efficiency and help find associations between morphological features and clinical outcome. For development of such prediction models, identifying invasive epithelial cells, and separating these from benign epithelial cells and in situ lesions would be the first step. In this study, we aimed to develop an AI model for segmentation of epithelial cells in sections from breast cancer. We generated epithelial ground truth masks by restaining hematoxylin and eosin (HE) sections with cytokeratin (CK) AE1/AE3, and by pathologists' annotations. HE/CK image pairs were used to train a convolutional neural network, and data augmentation was used to make the model more robust. Tissue microarrays (TMAs) from 839 patients, and whole slide images from two patients were used for training and evaluation of the models. The sections were derived from four cohorts of breast cancer patients. TMAs from 21 patients from a fifth cohort was used as a second test set. In quantitative evaluation, a mean Dice score of 0.70, 0.79, and 0.75 for invasive epithelial cells, benign epithelial cells, and in situ lesions, respectively, were achieved. In qualitative scoring (0-5) by pathologists, results were best for all epithelium and invasive epithelium, with scores of 4.7 and 4.4. Scores for benign epithelium and in situ lesions were 3.7 and 2.0. The proposed model segmented epithelial cells in HE stained breast cancer slides well, but further work is needed for accurate division between the classes. Immunohistochemistry, together with pathologists' annotations, enabled the creation of accurate ground truths. The model is made freely available in FastPathology and the code is available at https://github.com/AICAN-Research/breast-epithelium-segmentation
URLs: https://github.com/AICAN-Research/breast-epithelium-segmentation
Authors: Sai Karthikey Pentapati, Anshul Rai, Arkady Ten, Chaitanya Atluru, Alan Bovik
Abstract: High-resolution texture maps are necessary for representing real-world objects accurately with 3D meshes. The large sizes of textures can bottleneck the real-time rendering of high-quality virtual 3D scenes on devices having low computational budgets and limited memory. Downsampling the texture maps directly addresses the issue, albeit at the cost of visual fidelity. Traditionally, downsampling of texture maps is performed using methods like bicubic interpolation and the Lanczos algorithm. These methods ignore the geometric layout of the mesh and its UV parametrization and also do not account for the rendering process used to obtain the final visualization that the users will experience. Towards filling these gaps, we introduce GeoScaler, which is a method of downsampling texture maps of 3D meshes while incorporating geometric cues, and by maximizing the visual fidelity of the rendered views of the textured meshes. We show that the textures generated by GeoScaler deliver significantly better quality rendered images compared to those generated by traditional downsampling methods
Authors: Yilin Liu, Yunkui Pang, Jiang Li, Yong Chen, Pew-Thian Yap
Abstract: Untrained networks inspired by deep image prior have shown promising capabilities in recovering a high-quality image from noisy or partial measurements, without requiring training data. Their success has been widely attributed to the spectral bias acting as an implicit regularization induced by suitable network architectures. However, applications of such network-based priors often entail superfluous architectural decisions, overfitting risks, and slow optimization, all of which hinder their practicality. In this work, we propose efficient, architecture-agnostic methods for a more direct frequency control over the network priors: 1) constraining the bandwidth of the white-noise input, 2) controlling the bandwidth of the interpolation-based upsamplers, and 3) regularizing the Lipschitz constants of the layers. We show that even with just one extra line of code, the overfitting issues in underperforming architectures can be alleviated such that their performance gaps with the high-performing counterparts can be largely closed despite their distinct configurations, mitigating the need for architecture tuning. This then makes it possible to employ a more compact model to achieve similar or superior performance to larger models with greater efficiency. Our regularized network priors compare favorably with current supervised and self-supervised methods on MRI reconstruction and image inpainting tasks, serving as a stronger zero-shot baseline reconstructor. Our code will be made publicly available.
Authors: Ana-Maria Marcu, Long Chen, Jan H\"unermann, Alice Karnsund, Benoit Hanotte, Prajwal Chidananda, Saurabh Nair, Vijay Badrinarayanan, Alex Kendall, Jamie Shotton, Elahe Arani, Oleg Sinavski
Abstract: Autonomous driving has long faced a challenge with public acceptance due to the lack of explainability in the decision-making process. Video question-answering (QA) in natural language provides the opportunity for bridging this gap. Nonetheless, evaluating the performance of Video QA models has proved particularly tough due to the absence of comprehensive benchmarks. To fill this gap, we introduce LingoQA, a benchmark specifically for autonomous driving Video QA. The LingoQA trainable metric demonstrates a 0.95 Spearman correlation coefficient with human evaluations. We introduce a Video QA dataset of central London consisting of 419k samples that we release with the paper. We establish a baseline vision-language model and run extensive ablation studies to understand its performance.
Authors: Peng Gao, Peng Wang, Feng Gao, Fei Wang, Ruyue Yuan
Abstract: As a long-term vision in the field of artificial intelligence, the core goal of embodied intelligence is to improve the perception, understanding, and interaction capabilities of agents and the environment. Vision-language navigation (VLN), as a critical research path to achieve embodied intelligence, focuses on exploring how agents use natural language to communicate effectively with humans, receive and understand instructions, and ultimately rely on visual information to achieve accurate navigation. VLN integrates artificial intelligence, natural language processing, computer vision, and robotics. This field faces technical challenges but shows potential for application such as human-computer interaction. However, due to the complex process involved from language understanding to action execution, VLN faces the problem of aligning visual information and language instructions, improving generalization ability, and many other challenges. This survey systematically reviews the research progress of VLN and details the research direction of VLN with embodied intelligence. After a detailed summary of its system architecture and research based on methods and commonly used benchmark datasets, we comprehensively analyze the problems and challenges faced by current research and explore the future development direction of this field, aiming to provide a practical reference for researchers.
Authors: Xinjie Zhang, Shenyuan Gao, Zhening Liu, Jiawei Shao, Xingtong Ge, Dailan He, Tongda Xu, Yan Wang, Jun Zhang
Abstract: Existing learning-based stereo image codec adopt sophisticated transformation with simple entropy models derived from single image codecs to encode latent representations. However, those entropy models struggle to effectively capture the spatial-disparity characteristics inherent in stereo images, which leads to suboptimal rate-distortion results. In this paper, we propose a stereo image compression framework, named CAMSIC. CAMSIC independently transforms each image to latent representation and employs a powerful decoder-free Transformer entropy model to capture both spatial and disparity dependencies, by introducing a novel content-aware masked image modeling (MIM) technique. Our content-aware MIM facilitates efficient bidirectional interaction between prior information and estimated tokens, which naturally obviates the need for an extra Transformer decoder. Experiments show that our stereo image codec achieves state-of-the-art rate-distortion performance on two stereo image datasets Cityscapes and InStereo2K with fast encoding and decoding speed.
Authors: Yuxuan Wang, Xuanyu Yi, Zike Wu, Na Zhao, Long Chen, Hanwang Zhang
Abstract: The advent of 3D Gaussian Splatting (3DGS) has revolutionized 3D editing, offering efficient, high-fidelity rendering and enabling precise local manipulations. Currently, diffusion-based 2D editing models are harnessed to modify multi-view rendered images, which then guide the editing of 3DGS models. However, this approach faces a critical issue of multi-view inconsistency, where the guidance images exhibit significant discrepancies across views, leading to mode collapse and visual artifacts of 3DGS. To this end, we introduce View-consistent Editing (VcEdit), a novel framework that seamlessly incorporates 3DGS into image editing processes, ensuring multi-view consistency in edited guidance images and effectively mitigating mode collapse issues. VcEdit employs two innovative consistency modules: the Cross-attention Consistency Module and the Editing Consistency Module, both designed to reduce inconsistencies in edited images. By incorporating these consistency modules into an iterative pattern, VcEdit proficiently resolves the issue of multi-view inconsistency, facilitating high-quality 3DGS editing across a diverse range of scenes.