Authors: Zhiqiang Wang, Dejia Xu, Rana Muhammad Shahroz Khan, Yanbin Lin, Zhiwen Fan, Xingquan Zhu
Abstract: Image geolocation is a critical task in various image-understanding applications. However, existing methods often fail when analyzing challenging, in-the-wild images. Inspired by the exceptional background knowledge of multimodal language models, we systematically evaluate their geolocation capabilities using a novel image dataset and a comprehensive evaluation framework. We first collect images from various countries via Google Street View. Then, we conduct training-free and training-based evaluations on closed-source and open-source multi-modal language models. we conduct both training-free and training-based evaluations on closed-source and open-source multimodal language models. Our findings indicate that closed-source models demonstrate superior geolocation abilities, while open-source models can achieve comparable performance through fine-tuning.
Authors: Muhammad Zubair Irshad
Abstract: Recent advances in deep learning have led to a data-centric intelligence i.e. artificially intelligent models unlocking the potential to ingest a large amount of data and be really good at performing digital tasks such as text-to-image generation, machine-human conversation, and image recognition. This thesis covers the topic of learning with structured inductive bias and priors to design approaches and algorithms unlocking the potential of principle-centric intelligence. Prior knowledge (priors for short), often available in terms of past experience as well as assumptions of how the world works, helps the autonomous agent generalize better and adapt their behavior based on past experience. In this thesis, I demonstrate the use of prior knowledge in three different robotics perception problems. 1. object-centric 3D reconstruction, 2. vision and language for decision-making, and 3. 3D scene understanding. To solve these challenging problems, I propose various sources of prior knowledge including 1. geometry and appearance priors from synthetic data, 2. modularity and semantic map priors and 3. semantic, structural, and contextual priors. I study these priors for solving robotics 3D perception tasks and propose ways to efficiently encode them in deep learning models. Some priors are used to warm-start the network for transfer learning, others are used as hard constraints to restrict the action space of robotics agents. While classical techniques are brittle and fail to generalize to unseen scenarios and data-centric approaches require a large amount of labeled data, this thesis aims to build intelligent agents which require very-less real-world data or data acquired only from simulation to generalize to highly dynamic and cluttered environments in novel simulations (i.e. sim2sim) or real-world unseen environments (i.e. sim2real) for a holistic scene understanding of the 3D world.
Authors: Qi Zhang, Guohua Geng, Longquan Yan, Pengbo Zhou, Zhaodi Li, Kang Li, Qinglin Liu
Abstract: Diffusion models and multi-scale features are essential components in semantic segmentation tasks that deal with remote-sensing images. They contribute to improved segmentation boundaries and offer significant contextual information. U-net-like architectures are frequently employed in diffusion models for segmentation tasks. These architectural designs include dense skip connections that may pose challenges for interpreting intermediate features. Consequently, they might not efficiently convey semantic information throughout various layers of the encoder-decoder architecture. To address these challenges, we propose a new model for semantic segmentation known as the diffusion model with parallel multi-scale branches. This model consists of Parallel Multiscale Diffusion modules (P-MSDiff) and a Cross-Bridge Linear Attention mechanism (CBLA). P-MSDiff enhances the understanding of semantic information across multiple levels of granularity and detects repetitive distribution data through the integration of recursive denoising branches. It further facilitates the amalgamation of data by connecting relevant branches to the primary framework to enable concurrent denoising. Furthermore, within the interconnected transformer architecture, the LA module has been substituted with the CBLA module. This module integrates a semidefinite matrix linked to the query into the dot product computation of keys and values. This integration enables the adaptation of queries within the LA framework. This adjustment enhances the structure for multi-head attention computation, leading to enhanced network performance and CBLA is a plug-and-play module. Our model demonstrates superior performance based on the J1 metric on both the UAVid and Vaihingen Building datasets, showing improvements of 1.60% and 1.40% over strong baseline models, respectively.
Authors: Selim Kuzucu, Kemal Oksuz, Jonathan Sadeghi, Puneet K. Dokania
Abstract: Reliable usage of object detectors require them to be calibrated -- a crucial problem that requires careful attention. Recent approaches towards this involve (1) designing new loss functions to obtain calibrated detectors by training them from scratch, and (2) post-hoc Temperature Scaling (TS) that learns to scale the likelihood of a trained detector to output calibrated predictions. These approaches are then evaluated based on a combination of Detection Expected Calibration Error (D-ECE) and Average Precision. In this work, via extensive analysis and insights, we highlight that these recent evaluation frameworks, evaluation metrics, and the use of TS have notable drawbacks leading to incorrect conclusions. As a step towards fixing these issues, we propose a principled evaluation framework to jointly measure calibration and accuracy of object detectors. We also tailor efficient and easy-to-use post-hoc calibration approaches such as Platt Scaling and Isotonic Regression specifically for object detection task. Contrary to the common notion, our experiments show that once designed and evaluated properly, post-hoc calibrators, which are extremely cheap to build and use, are much more powerful and effective than the recent train-time calibration methods. To illustrate, D-DETR with our post-hoc Isotonic Regression calibrator outperforms the recent train-time state-of-the-art calibration method Cal-DETR by more than 7 D-ECE on the COCO dataset. Additionally, we propose improved versions of the recently proposed Localization-aware ECE and show the efficacy of our method on these metrics as well. Code is available at: https://github.com/fiveai/detection_calibration.
Authors: Yi Wang, Conrad M Albrecht, Xiao Xiang Zhu
Abstract: Self-supervised pretraining on large-scale satellite data has raised great interest in building Earth observation (EO) foundation models. However, many important resources beyond pure satellite imagery, such as land-cover-land-use products that provide free global semantic information, as well as vision foundation models that hold strong knowledge of the natural world, tend to be overlooked. In this work, we show these free additional resources not only help resolve common contrastive learning bottlenecks, but also significantly boost the efficiency and effectiveness of EO pretraining. Specifically, we first propose soft contrastive learning that optimizes cross-scene soft similarity based on land-cover-generated multi-label supervision, naturally solving the issue of multiple positive samples and too strict positive matching in complex scenes. Second, we explore cross-domain continual pretraining for both multispectral and SAR imagery, building efficient EO foundation models from strongest vision models such as DINOv2. Integrating simple weight-initialization and Siamese masking strategies into our soft contrastive learning framework, we demonstrate impressive continual pretraining performance even when the input channels and modalities are not aligned. Without prohibitive training, we produce multispectral and SAR foundation models that achieve significantly better results in 9 out of 10 downstream tasks than most existing SOTA models. For example, our ResNet50/ViT-S achieve 84.8/85.0 linear probing mAP scores on BigEarthNet-10\% which are better than most existing ViT-L models; under the same setting, our ViT-B sets a new record of 86.8 in multispectral, and 82.5 in SAR, the latter even better than many multispectral models. Dataset and models are available at https://github.com/zhu-xlab/softcon.
Authors: Serdar Yildiz, Ahmet Nezih Kasim
Abstract: The growing importance of person reidentification in computer vision has highlighted the need for more extensive and diverse datasets. In response, we introduce the ENTIRe-ID dataset, an extensive collection comprising over 4.45 million images from 37 different cameras in varied environments. This dataset is uniquely designed to tackle the challenges of domain variability and model generalization, areas where existing datasets for person re-identification have fallen short. The ENTIRe-ID dataset stands out for its coverage of a wide array of real-world scenarios, encompassing various lighting conditions, angles of view, and diverse human activities. This design ensures a realistic and robust training platform for ReID models. The ENTIRe-ID dataset is publicly available at https://serdaryildiz.github.io/ENTIRe-ID
Authors: Krishnakant Singh, Thanush Navaratnam, Jannik Holmer, Simone Schaub-Meyer, Stefan Roth
Abstract: A long-standing challenge in developing machine learning approaches has been the lack of high-quality labeled data. Recently, models trained with purely synthetic data, here termed synthetic clones, generated using large-scale pre-trained diffusion models have shown promising results in overcoming this annotation bottleneck. As these synthetic clone models progress, they are likely to be deployed in challenging real-world settings, yet their suitability remains understudied. Our work addresses this gap by providing the first benchmark for three classes of synthetic clone models, namely supervised, self-supervised, and multi-modal ones, across a range of robustness measures. We show that existing synthetic self-supervised and multi-modal clones are comparable to or outperform state-of-the-art real-image baselines for a range of robustness metrics - shape bias, background bias, calibration, etc. However, we also find that synthetic clones are much more susceptible to adversarial and real-world noise than models trained with real data. To address this, we find that combining both real and synthetic data further increases the robustness, and that the choice of prompt used for generating synthetic images plays an important part in the robustness of synthetic clones.
Authors: Hao Chen, Yujin Han, Diganta Misra, Xiang Li, Kai Hu, Difan Zou, Masashi Sugiyama, Jindong Wang, Bhiksha Raj
Abstract: Diffusion models (DMs) have shown remarkable capabilities in generating realistic high-quality images, audios, and videos. They benefit significantly from extensive pre-training on large-scale datasets, including web-crawled data with paired data and conditions, such as image-text and image-class pairs. Despite rigorous filtering, these pre-training datasets often inevitably contain corrupted pairs where conditions do not accurately describe the data. This paper presents the first comprehensive study on the impact of such corruption in pre-training data of DMs. We synthetically corrupt ImageNet-1K and CC3M to pre-train and evaluate over 50 conditional DMs. Our empirical findings reveal that various types of slight corruption in pre-training can significantly enhance the quality, diversity, and fidelity of the generated images across different DMs, both during pre-training and downstream adaptation stages. Theoretically, we consider a Gaussian mixture model and prove that slight corruption in the condition leads to higher entropy and a reduced 2-Wasserstein distance to the ground truth of the data distribution generated by the corruptly trained DMs. Inspired by our analysis, we propose a simple method to improve the training of DMs on practical datasets by adding condition embedding perturbations (CEP). CEP significantly improves the performance of various DMs in both pre-training and downstream tasks. We hope that our study provides new insights into understanding the data and pre-training processes of DMs.
Authors: Minghao Guo, Bohan Wang, Pingchuan Ma, Tianyuan Zhang, Crystal Elaine Owens, Chuang Gan, Joshua B. Tenenbaum, Kaiming He, Wojciech Matusik
Abstract: We present a computational framework that transforms single images into 3D physical objects. The visual geometry of a physical object in an image is determined by three orthogonal attributes: mechanical properties, external forces, and rest-shape geometry. Existing single-view 3D reconstruction methods often overlook this underlying composition, presuming rigidity or neglecting external forces. Consequently, the reconstructed objects fail to withstand real-world physical forces, resulting in instability or undesirable deformation -- diverging from their intended designs as depicted in the image. Our optimization framework addresses this by embedding physical compatibility into the reconstruction process. We explicitly decompose the three physical attributes and link them through static equilibrium, which serves as a hard constraint, ensuring that the optimized physical shapes exhibit desired physical behaviors. Evaluations on a dataset collected from Objaverse demonstrate that our framework consistently enhances the physical realism of 3D models over existing methods. The utility of our framework extends to practical applications in dynamic simulations and 3D printing, where adherence to physical compatibility is paramount.
Authors: Yisu Liu, Jinyang An, Wanqian Zhang, Dayan Wu, Jingzi Gu, Zheng Lin, Weiping Wang
Abstract: With the development of diffusion-based customization methods like DreamBooth, individuals now have access to train the models that can generate their personalized images. Despite the convenience, malicious users have misused these techniques to create fake images, thereby triggering a privacy security crisis. In light of this, proactive adversarial attacks are proposed to protect users against customization. The adversarial examples are trained to distort the customization model's outputs and thus block the misuse. In this paper, we propose DisDiff (Disrupting Diffusion), a novel adversarial attack method to disrupt the diffusion model outputs. We first delve into the intrinsic image-text relationships, well-known as cross-attention, and empirically find that the subject-identifier token plays an important role in guiding image generation. Thus, we propose the Cross-Attention Erasure module to explicitly "erase" the indicated attention maps and disrupt the text guidance. Besides,we analyze the influence of the sampling process of the diffusion model on Projected Gradient Descent (PGD) attack and introduce a novel Merit Sampling Scheduler to adaptively modulate the perturbation updating amplitude in a step-aware manner. Our DisDiff outperforms the state-of-the-art methods by 12.75% of FDFR scores and 7.25% of ISM scores across two facial benchmarks and two commonly used prompts on average.
Authors: Jiachen Liang, Ruibing Hou, Hong Chang, Bingpeng Ma, Shiguang Shan, Xilin Chen
Abstract: Traditional semi-supervised learning (SSL) assumes that the feature distributions of labeled and unlabeled data are consistent which rarely holds in realistic scenarios. In this paper, we propose a novel SSL setting, where unlabeled samples are drawn from a mixed distribution that deviates from the feature distribution of labeled samples. Under this setting, previous SSL methods tend to predict wrong pseudo-labels with the model fitted on labeled data, resulting in noise accumulation. To tackle this issue, we propose Self-Supervised Feature Adaptation (SSFA), a generic framework for improving SSL performance when labeled and unlabeled data come from different distributions. SSFA decouples the prediction of pseudo-labels from the current model to improve the quality of pseudo-labels. Particularly, SSFA incorporates a self-supervised task into the SSL framework and uses it to adapt the feature extractor of the model to the unlabeled data. In this way, the extracted features better fit the distribution of unlabeled data, thereby generating high-quality pseudo-labels. Extensive experiments show that our proposed SSFA is applicable to various pseudo-label-based SSL learners and significantly improves performance in labeled, unlabeled, and even unseen distributions.
Authors: Yang Chen, Tian He, Junfeng Fu, Ling Wang, Jingcai Guo, Hong Cheng
Abstract: Supervised and self-supervised learning are two main training paradigms for skeleton-based human action recognition. However, the former one-hot classification requires labor-intensive predefined action categories annotations, while the latter involves skeleton transformations (e.g., cropping) in the pretext tasks that may impair the skeleton structure. To address these challenges, we introduce a novel skeleton-based training framework (C$^2$VL) based on Cross-modal Contrastive learning that uses the progressive distillation to learn task-agnostic human skeleton action representation from the Vision-Language knowledge prompts. Specifically, we establish the vision-language action concept space through vision-language knowledge prompts generated by pre-trained large multimodal models (LMMs), which enrich the fine-grained details that the skeleton action space lacks. Moreover, we propose the intra-modal self-similarity and inter-modal cross-consistency softened targets in the cross-modal contrastive process to progressively control and guide the degree of pulling vision-language knowledge prompts and corresponding skeletons closer. These soft instance discrimination and self-knowledge distillation strategies contribute to the learning of better skeleton-based action representations from the noisy skeleton-vision-language pairs. During the inference phase, our method requires only the skeleton data as the input for action recognition and no longer for vision-language prompts. Extensive experiments show that our method achieves state-of-the-art results on NTU RGB+D 60, NTU RGB+D 120, and PKU-MMD datasets. The code will be available in the future.
Authors: Yuanjiang Luo, Hongxiang Li, Xuan Wu, Meng Cao, Xiaoshuang Huang, Zhihong Zhu, Peixi Liao, Hu Chen, Yi Zhang
Abstract: Existing mainstream approaches follow the encoder-decoder paradigm for generating radiology reports. They focus on improving the network structure of encoders and decoders, which leads to two shortcomings: overlooking the modality gap and ignoring report content constraints. In this paper, we proposed Textual Inversion and Self-supervised Refinement (TISR) to address the above two issues. Specifically, textual inversion can project text and image into the same space by representing images as pseudo words to eliminate the cross-modeling gap. Subsequently, self-supervised refinement refines these pseudo words through contrastive loss computation between images and texts, enhancing the fidelity of generated reports to images. Notably, TISR is orthogonal to most existing methods, plug-and-play. We conduct experiments on two widely-used public datasets and achieve significant improvements on various baselines, which demonstrates the effectiveness and generalization of TISR. The code will be available soon.
Authors: Wooseok Shin, Hyun Joon Park, Jin Sob Kim, Sung Won Han
Abstract: In semi-supervised semantic segmentation, the Mean Teacher- and co-training-based approaches are employed to mitigate confirmation bias and coupling problems. However, despite their high performance, these approaches frequently involve complex training pipelines and a substantial computational burden, limiting the scalability and compatibility of these methods. In this paper, we propose a PrevMatch framework that effectively mitigates the aforementioned limitations by maximizing the utilization of the temporal knowledge obtained during the training process. The PrevMatch framework relies on two core strategies: (1) we reconsider the use of temporal knowledge and thus directly utilize previous models obtained during training to generate additional pseudo-label guidance, referred to as previous guidance. (2) we design a highly randomized ensemble strategy to maximize the effectiveness of the previous guidance. Experimental results on four benchmark semantic segmentation datasets confirm that the proposed method consistently outperforms existing methods across various evaluation protocols. In particular, with DeepLabV3+ and ResNet-101 network settings, PrevMatch outperforms the existing state-of-the-art method, Diverse Co-training, by +1.6 mIoU on Pascal VOC with only 92 annotated images, while achieving 2.4 times faster training. Furthermore, the results indicate that PrevMatch induces stable optimization, particularly in benefiting classes that exhibit poor performance. Code is available at https://github.com/wooseok-shin/PrevMatch
Authors: Junming Ren, Zhoujian Xiao, Yujia Zhang, Yujie Yang, Ling He, Ezra Yoon, Stephen Temitayo Bello, Xi Chen, Dapeng Wu, Micky Tortorella, Jufang He
Abstract: In the preclinical translational studies, drug candidates with remarkable anti-epileptic efficacy demonstrate long-term suppression of spontaneous recurrent seizures (SRSs), particularly convulsive seizures (CSs), in mouse models of chronic epilepsy. However, the current methods for monitoring CSs have limitations in terms of invasiveness, specific laboratory settings, high cost, and complex operation, which hinder drug screening efforts. In this study, a camera-based system for automated detection of CSs in chronically epileptic mice is first established to screen potential anti-epilepsy drugs.
Authors: Jing Xu, Anqi Zhu, Jingyu Lin, Qiuhong Ke, Cunjian Chen
Abstract: Human action recognition is a crucial task in computer vision systems. However, in real-world scenarios, human actions often fall outside the distribution of training data, requiring a model to both recognize in-distribution (ID) actions and reject out-of-distribution (OOD) ones. Despite its importance, there has been limited research on OOD detection in human actions. Existing works on OOD detection mainly focus on image data with RGB structure, and many methods are post-hoc in nature. While these methods are convenient and computationally efficient, they often lack sufficient accuracy and fail to consider the presence of OOD samples. To address these challenges, we propose a novel end-to-end skeleton-based model called Action-OOD, specifically designed for OOD human action detection. Unlike some existing approaches that may require prior knowledge of existing OOD data distribution, our model solely utilizes in-distribution (ID) data during the training stage, effectively mitigating the overconfidence issue prevalent in OOD detection. We introduce an attention-based feature fusion block, which enhances the model's capability to recognize unknown classes while preserving classification accuracy for known classes. Further, we present a novel energy-based loss function and successfully integrate it with the traditional cross-entropy loss to maximize the separation of data distributions between ID and OOD. Through extensive experiments conducted on NTU-RGB+D 60, NTU-RGB+D 120, and Kinetics-400 datasets, we demonstrate the superior performance of our proposed approach compared to state-of-the-art methods. Our findings underscore the effectiveness of classic OOD detection techniques in the context of skeleton-based action recognition tasks, offering promising avenues for future research in this field. Code will be available at: https://github.com/YilliaJing/Action-OOD.git.
Authors: Nerea Aranjuelo, Siyu Huang, Ignacio Arganda-Carreras, Luis Unzueta, Oihana Otaegui, Hanspeter Pfister, Donglai Wei
Abstract: Gaze-annotated facial data is crucial for training deep neural networks (DNNs) for gaze estimation. However, obtaining these data is labor-intensive and requires specialized equipment due to the challenge of accurately annotating the gaze direction of a subject. In this work, we present a generative framework to create annotated gaze data by leveraging the benefits of labeled and unlabeled data sources. We propose a Gaze-aware Compositional GAN that learns to generate annotated facial images from a limited labeled dataset. Then we transfer this model to an unlabeled data domain to take advantage of the diversity it provides. Experiments demonstrate our approach's effectiveness in generating within-domain image augmentations in the ETH-XGaze dataset and cross-domain augmentations in the CelebAMask-HQ dataset domain for gaze estimation DNN training. We also show additional applications of our work, which include facial image editing and gaze redirection.
Authors: Richard Luo, Austin Peng, Adithya Vasudev, Rishabh Jain
Abstract: Video is an increasingly prominent and information-dense medium, yet it poses substantial challenges for language models. A typical video consists of a sequence of shorter segments, or shots, that collectively form a coherent narrative. Each shot is analogous to a word in a sentence where multiple data streams of information (such as visual and auditory data) must be processed simultaneously. Comprehension of the entire video requires not only understanding the visual-audio information of each shot but also requires that the model links the ideas between each shot to generate a larger, all-encompassing story. Despite significant progress in the field, current works often overlook videos' more granular shot-by-shot semantic information. In this project, we propose a family of efficient large language vision models (LLVMs) to boost video summarization and captioning called Shotluck Holmes. By leveraging better pretraining and data collection strategies, we extend the abilities of existing small LLVMs from being able to understand a picture to being able to understand a sequence of frames. Specifically, we show that Shotluck Holmes achieves better performance than state-of-the-art results on the Shot2Story video captioning and summary task with significantly smaller and more computationally efficient models.
Authors: Hansang Lee, Haeil Lee, Helen Hong
Abstract: In this paper, we propose a novel data augmentation technique called GenMix, which combines generative and mixture approaches to leverage the strengths of both methods. While generative models excel at creating new data patterns, they face challenges such as mode collapse in GANs and difficulties in training diffusion models, especially with limited medical imaging data. On the other hand, mixture models enhance class boundary regions but tend to favor the major class in scenarios with class imbalance. To address these limitations, GenMix integrates both approaches to complement each other. GenMix operates in two stages: (1) training a generative model to produce synthetic images, and (2) performing mixup between synthetic and real data. This process improves the quality and diversity of synthetic data while simultaneously benefiting from the new pattern learning of generative models and the boundary enhancement of mixture models. We validate the effectiveness of our method on the task of classifying focal liver lesions (FLLs) in CT images. Our results demonstrate that GenMix enhances the performance of various generative models, including DCGAN, StyleGAN, Textual Inversion, and Diffusion Models. Notably, the proposed method with Textual Inversion outperforms other methods without fine-tuning diffusion model on the FLL dataset.
Authors: Weichao Zhao, Hezhen Hu, Wengang Zhou, Yunyao Mao, Min Wang, Houqiang Li
Abstract: Sign language recognition (SLR) has long been plagued by insufficient model representation capabilities. Although current pre-training approaches have alleviated this dilemma to some extent and yielded promising performance by employing various pretext tasks on sign pose data, these methods still suffer from two primary limitations: 1) Explicit motion information is usually disregarded in previous pretext tasks, leading to partial information loss and limited representation capability. 2) Previous methods focus on the local context of a sign pose sequence, without incorporating the guidance of the global meaning of lexical signs. To this end, we propose a Motion-Aware masked autoencoder with Semantic Alignment (MASA) that integrates rich motion cues and global semantic information in a self-supervised learning paradigm for SLR. Our framework contains two crucial components, i.e., a motion-aware masked autoencoder (MA) and a momentum semantic alignment module (SA). Specifically, in MA, we introduce an autoencoder architecture with a motion-aware masked strategy to reconstruct motion residuals of masked frames, thereby explicitly exploring dynamic motion cues among sign pose sequences. Moreover, in SA, we embed our framework with global semantic awareness by aligning the embeddings of different augmented samples from the input sequence in the shared latent space. In this way, our framework can simultaneously learn local motion cues and global semantic features for comprehensive sign language representation. Furthermore, we conduct extensive experiments to validate the effectiveness of our method, achieving new state-of-the-art performance on four public benchmarks.
Authors: Shuzhou Yang, Yu Wang, Haijie Li, Jiarui Meng, Xiandong Meng, Jian Zhang
Abstract: Single image-to-3D generation is pivotal for crafting controllable 3D assets. Given its underconstrained nature, we leverage geometric priors from a 3D novel view generation diffusion model and appearance priors from a 2D image generation method to guide the optimization process. We note that a disparity exists between the training datasets of 2D and 3D diffusion models, leading to their outputs showing marked differences in appearance. Specifically, 2D models tend to deliver more detailed visuals, whereas 3D models produce consistent yet over-smooth results across different views. Hence, we optimize a set of 3D Gaussians using 3D priors in spatial domain to ensure geometric consistency, while exploiting 2D priors in the frequency domain through Fourier transform for higher visual quality. This 2D-3D hybrid Fourier Score Distillation objective function (dubbed hy-FSD), can be integrated into existing 3D generation methods, yielding significant performance improvements. With this technique, we further develop an image-to-3D generation pipeline to create high-quality 3D objects within one minute, named Fourier123. Extensive experiments demonstrate that Fourier123 excels in efficient generation with rapid convergence speed and visual-friendly generation results.
Authors: Davide Coppola, Hwee Kuan Lee
Abstract: This study explores the impact of adversarial perturbations on Convolutional Neural Networks (CNNs) with the aim of enhancing the understanding of their underlying mechanisms. Despite numerous defense methods proposed in the literature, there is still an incomplete understanding of this phenomenon. Instead of treating the entire model as vulnerable, we propose that specific feature maps learned during training contribute to the overall vulnerability. To investigate how the hidden representations learned by a CNN affect its vulnerability, we introduce the Adversarial Intervention framework. Experiments were conducted on models trained on three well-known computer vision datasets, subjecting them to attacks of different nature. Our focus centers on the effects that adversarial perturbations to a model's initial layer have on the overall behavior of the model. Empirical results revealed compelling insights: a) perturbing selected channel combinations in shallow layers causes significant disruptions; b) the channel combinations most responsible for the disruptions are common among different types of attacks; c) despite shared vulnerable combinations of channels, different attacks affect hidden representations with varying magnitudes; d) there exists a positive correlation between a kernel's magnitude and its vulnerability. In conclusion, this work introduces a novel framework to study the vulnerability of a CNN model to adversarial perturbations, revealing insights that contribute to a deeper understanding of the phenomenon. The identified properties pave the way for the development of efficient ad-hoc defense mechanisms in future applications.
Authors: Haiyu Zhang, Xinyuan Chen, Yaohui Wang, Xihui Liu, Yunhong Wang, Yu Qiao
Abstract: Current 4D generation methods have achieved noteworthy efficacy with the aid of advanced diffusion generative models. However, these methods lack multi-view spatial-temporal modeling and encounter challenges in integrating diverse prior knowledge from multiple diffusion models, resulting in inconsistent temporal appearance and flickers. In this paper, we propose a novel 4D generation pipeline, namely 4Diffusion aimed at generating spatial-temporally consistent 4D content from a monocular video. We first design a unified diffusion model tailored for multi-view video generation by incorporating a learnable motion module into a frozen 3D-aware diffusion model to capture multi-view spatial-temporal correlations. After training on a curated dataset, our diffusion model acquires reasonable temporal consistency and inherently preserves the generalizability and spatial consistency of the 3D-aware diffusion model. Subsequently, we propose 4D-aware Score Distillation Sampling loss, which is based on our multi-view video diffusion model, to optimize 4D representation parameterized by dynamic NeRF. This aims to eliminate discrepancies arising from multiple diffusion models, allowing for generating spatial-temporally consistent 4D content. Moreover, we devise an anchor loss to enhance the appearance details and facilitate the learning of dynamic NeRF. Extensive qualitative and quantitative experiments demonstrate that our method achieves superior performance compared to previous methods.
Authors: Kidist Amde Mekonnen, Nicola Dall'Asen, Paolo Rota
Abstract: Diffusion Probabilistic Models (DPMs) have emerged as a powerful class of deep generative models, achieving remarkable performance in image synthesis tasks. However, these models face challenges in terms of widespread adoption due to their reliance on sequential denoising steps during sample generation. This dependence leads to substantial computational requirements, making them unsuitable for resource-constrained or real-time processing systems. To address these challenges, we propose a novel method that integrates denoising phases directly into the model's architecture, thereby reducing the need for resource-intensive computations. Our approach combines diffusion models with generative adversarial networks (GANs) through knowledge distillation, enabling more efficient training and evaluation. By utilizing a pre-trained diffusion model as a teacher model, we train a student model through adversarial learning, employing layerwise transformations for denoising and submodules for predicting the teacher model's output at various points in time. This integration significantly reduces the number of parameters and denoising steps required, leading to improved sampling speed at test time. We validate our method with extensive experiments, demonstrating comparable performance with reduced computational requirements compared to existing approaches. By enabling the deployment of diffusion models on resource-constrained devices, our research mitigates their computational burden and paves the way for wider accessibility and practical use across the research community and end-users. Our code is publicly available at https://github.com/kidist-amde/Adv-KD
Authors: Kidist Amde Mekonnen
Abstract: Deep learning algorithms have a large number of trainable parameters often with sizes of hundreds of thousands or more. Training this algorithm requires a large amount of training data and generating a sufficiently large dataset for these algorithms is costly\cite{noguchi2019image}. GANs are generative neural networks that use two deep learning networks that are competing with each other. The networks are generator and discriminator networks. The generator tries to generate realistic images which resemble the actual training dataset by approximating the training data distribution and the discriminator is trained to classify images as real or fake(generated)\cite{goodfellow2016nips}. Training these GAN algorithms also requires a large amount of training dataset\cite{noguchi2019image}. In this study, the aim is to address the question, "Given an unconditioned pretrained generator network and a pretrained classifier, is it feasible to develop a conditioned generator without relying on any training dataset?" The paper begins with a general introduction to the problem. The subsequent sections are structured as follows: Section 2 provides background information on the problem. Section 3 reviews relevant literature on the topic. Section 4 outlines the methodology employed in this study. Section 5 presents the experimental results. Section 6 discusses the findings and proposes potential future research directions. Finally, Section 7 offers concluding remarks. The implementation can be accessed \href{https://github.com/kidist-amde/BigGAN-PyTorch}{here}.
Authors: Zhaorui Tan, Chengrui Zhang, Xi Yang, Jie Sun, Kaizhu Huang
Abstract: Generalized category discovery presents a challenge in a realistic scenario, which requires the model's generalization ability to recognize unlabeled samples from known and unknown categories. This paper revisits the challenge of generalized category discovery through the lens of information maximization (InfoMax) with a probabilistic parametric classifier. Our findings reveal that ensuring independence between known and unknown classes while concurrently assuming a uniform probability distribution across all classes, yields an enlarged margin among known and unknown classes that promotes the model's performance. To achieve the aforementioned independence, we propose a novel InfoMax-based method, Regularized Parametric InfoMax (RPIM), which adopts pseudo labels to supervise unlabeled samples during InfoMax, while proposing a regularization to ensure the quality of the pseudo labels. Additionally, we introduce novel semantic-bias transformation to refine the features from the pre-trained model instead of direct fine-tuning to rescue the computational costs. Extensive experiments on six benchmark datasets validate the effectiveness of our method. RPIM significantly improves the performance regarding unknown classes, surpassing the state-of-the-art method by an average margin of 3.5%.
Authors: Takaya Tanaka, Yutaka Yamaguti
Abstract: Successive image generation using cyclic transformations is demonstrated by extending the CycleGAN model to transform images among three different categories. Repeated application of the trained generators produces sequences of images that transition among the different categories. The generated image sequences occupy a more limited region of the image space compared with the original training dataset. Quantitative evaluation using precision and recall metrics indicates that the generated images have high quality but reduced diversity relative to the training dataset. Such successive generation processes are characterized as chaotic dynamics in terms of dynamical system theory. Positive Lyapunov exponents estimated from the generated trajectories confirm the presence of chaotic dynamics, with the Lyapunov dimension of the attractor found to be comparable to the intrinsic dimension of the training data manifold. The results suggest that chaotic dynamics in the image space defined by the deep generative model contribute to the diversity of the generated images, constituting a novel approach for multi-class image generation. This model can be interpreted as an extension of classical associative memory to perform hetero-association among image categories.
Authors: Jin-Hee Lee, Jae-Keun Lee, Je-Seok Kim, Soon Kwon
Abstract: To ensure safe urban driving for autonomous platforms, it is crucial not only to develop high-performance object detection techniques but also to establish a diverse and representative dataset that captures various urban environments and object characteristics. To address these two issues, we have constructed a multi-class 3D LiDAR dataset reflecting diverse urban environments and object characteristics, and developed a robust 3D semi-supervised object detection (SSOD) based on a multiple teachers framework. This SSOD framework categorizes similar classes and assigns specialized teachers to each category. Through collaborative supervision among these category-specialized teachers, the student network becomes increasingly proficient, leading to a highly effective object detector. We propose a simple yet effective augmentation technique, Pie-based Point Compensating Augmentation (PieAug), to enable the teacher network to generate high-quality pseudo-labels. Extensive experiments on the WOD, KITTI, and our datasets validate the effectiveness of our proposed method and the quality of our dataset. Experimental results demonstrate that our approach consistently outperforms existing state-of-the-art 3D semi-supervised object detection methods across all datasets. We plan to release our multi-class LiDAR dataset and the source code available on our Github repository in the near future.
Authors: Yufei Wang, Zhihao Li, Lanqing Guo, Wenhan Yang, Alex C. Kot, Bihan Wen
Abstract: Recently, 3D Gaussian Splatting (3DGS) has become a promising framework for novel view synthesis, offering fast rendering speeds and high fidelity. However, the large number of Gaussians and their associated attributes require effective compression techniques. Existing methods primarily compress neural Gaussians individually and independently, i.e., coding all the neural Gaussians at the same time, with little design for their interactions and spatial dependence. Inspired by the effectiveness of the context model in image compression, we propose the first autoregressive model at the anchor level for 3DGS compression in this work. We divide anchors into different levels and the anchors that are not coded yet can be predicted based on the already coded ones in all the coarser levels, leading to more accurate modeling and higher coding efficiency. To further improve the efficiency of entropy coding, e.g., to code the coarsest level with no already coded anchors, we propose to introduce a low-dimensional quantized feature as the hyperprior for each anchor, which can be effectively compressed. Our work pioneers the context model in the anchor level for 3DGS representation, yielding an impressive size reduction of over 100 times compared to vanilla 3DGS and 15 times compared to the most recent state-of-the-art work Scaffold-GS, while achieving comparable or even higher rendering quality.
Authors: Hyeonjun Lee, Sehyun Hwang, Suha Kwak
Abstract: This paper introduces a novel approach to learning instance segmentation using extreme points, i.e., the topmost, leftmost, bottommost, and rightmost points, of each object. These points are readily available in the modern bounding box annotation process while offering strong clues for precise segmentation, and thus allows to improve performance at the same annotation cost with box-supervised methods. Our work considers extreme points as a part of the true instance mask and propagates them to identify potential foreground and background points, which are all together used for training a pseudo label generator. Then pseudo labels given by the generator are in turn used for supervised learning of our final model. On three public benchmarks, our method significantly outperforms existing box-supervised methods, further narrowing the gap with its fully supervised counterpart. In particular, our model generates high-quality masks when a target object is separated into multiple parts, where previous box-supervised methods often fail.
Authors: Mansi Kakkar, Dattesh Shanbhag, Chandan Aladahalli, Gurunath Reddy M
Abstract: Vision-language models have emerged as a powerful tool for previously challenging multi-modal classification problem in the medical domain. This development has led to the exploration of automated image description generation for multi-modal clinical scans, particularly for radiology report generation. Existing research has focused on clinical descriptions for specific modalities or body regions, leaving a gap for a model providing entire-body multi-modal descriptions. In this paper, we address this gap by automating the generation of standardized body station(s) and list of organ(s) across the whole body in multi-modal MR and CT radiological images. Leveraging the versatility of the Contrastive Language-Image Pre-training (CLIP), we refine and augment the existing approach through multiple experiments, including baseline model fine-tuning, adding station(s) as a superset for better correlation between organs, along with image and language augmentations. Our proposed approach demonstrates 47.6% performance improvement over baseline PubMedCLIP.
Authors: Riccardo Benaglia, Angelo Porrello, Pietro Buzzega, Simone Calderara, Rita Cucchiara
Abstract: Trajectory forecasting is crucial for video surveillance analytics, as it enables the anticipation of future movements for a set of agents, e.g. basketball players engaged in intricate interactions with long-term intentions. Deep generative models offer a natural learning approach for trajectory forecasting, yet they encounter difficulties in achieving an optimal balance between sampling fidelity and diversity. We address this challenge by leveraging Vector Quantized Variational Autoencoders (VQ-VAEs), which utilize a discrete latent space to tackle the issue of posterior collapse. Specifically, we introduce an instance-based codebook that allows tailored latent representations for each example. In a nutshell, the rows of the codebook are dynamically adjusted to reflect contextual information (i.e., past motion patterns extracted from the observed trajectories). In this way, the discretization process gains flexibility, leading to improved reconstructions. Notably, instance-level dynamics are injected into the codebook through low-rank updates, which restrict the customization of the codebook to a lower dimension space. The resulting discrete space serves as the basis of the subsequent step, which regards the training of a diffusion-based predictive model. We show that such a two-fold framework, augmented with instance-level discretization, leads to accurate and diverse forecasts, yielding state-of-the-art performance on three established benchmarks.
Authors: Bowen Zheng, Tianming Yang
Abstract: Diffusion Models (DMs) have achieved great success in image generation and other fields. By fine sampling through the trajectory defined by the SDE/ODE solver based on a well-trained score model, DMs can generate remarkable high-quality results. However, this precise sampling often requires multiple steps and is computationally demanding. To address this problem, instance-based distillation methods have been proposed to distill a one-step generator from a DM by having a simpler student model mimic a more complex teacher model. Yet, our research reveals an inherent limitations in these methods: the teacher model, with more steps and more parameters, occupies different local minima compared to the student model, leading to suboptimal performance when the student model attempts to replicate the teacher. To avoid this problem, we introduce a novel distributional distillation method, which uses an exclusive distributional loss. This method exceeds state-of-the-art (SOTA) results while requiring significantly fewer training images. Additionally, we show that DMs' layers are activated differently at different time steps, leading to an inherent capability to generate images in a single step. Freezing most of the convolutional layers in a DM during distributional distillation leads to further performance improvements. Our method achieves the SOTA results on CIFAR-10 (FID 1.54), AFHQv2 64x64 (FID 1.23), FFHQ 64x64 (FID 0.85) and ImageNet 64x64 (FID 1.16) with great efficiency. Most of those results are obtained with only 5 million training images within 6 hours on 8 A100 GPUs. This breakthrough not only enhances the understanding of efficient image generation models but also offers a scalable framework for advancing the state of the art in various applications.
Authors: Zhiming Meng, Hui Li, Zeyang Zhang, Zhongwei Shen, Yunlong Yu, Xiaoning Song, Xiaojun Wu
Abstract: Generative models are widely utilized to model the distribution of fused images in the field of infrared and visible image fusion. However, current generative models based fusion methods often suffer from unstable training and slow inference speed. To tackle this problem, a novel fusion method based on consistency model is proposed, termed as CoMoFusion, which can generate the high-quality images and achieve fast image inference speed. In specific, the consistency model is used to construct multi-modal joint features in the latent space with the forward and reverse process. Then, the infrared and visible features extracted by the trained consistency model are fed into fusion module to generate the final fused image. In order to enhance the texture and salient information of fused images, a novel loss based on pixel value selection is also designed. Extensive experiments on public datasets illustrate that our method obtains the SOTA fusion performance compared with the existing fusion methods.
Authors: Han Feng, Wenchao Ma, Quankai Gao, Xianwei Zheng, Nan Xue, Huijuan Xu
Abstract: Estimating 3D full-body avatars from AR/VR devices is essential for creating immersive experiences in AR/VR applications. This task is challenging due to the limited input from Head Mounted Devices, which capture only sparse observations from the head and hands. Predicting the full-body avatars, particularly the lower body, from these sparse observations presents significant difficulties. In this paper, we are inspired by the inherent property of the kinematic tree defined in the Skinned Multi-Person Linear (SMPL) model, where the upper body and lower body share only one common ancestor node, bringing the potential of decoupled reconstruction. We propose a stratified approach to decouple the conventional full-body avatar reconstruction pipeline into two stages, with the reconstruction of the upper body first and a subsequent reconstruction of the lower body conditioned on the previous stage. To implement this straightforward idea, we leverage the latent diffusion model as a powerful probabilistic generator, and train it to follow the latent distribution of decoupled motions explored by a VQ-VAE encoder-decoder model. Extensive experiments on AMASS mocap dataset demonstrate our state-of-the-art performance in the reconstruction of full-body motions.
Authors: Yumeng He, Yunbo Wang, Xiaokang Yang
Abstract: Decoupling the illumination in 3D scenes is crucial for novel view synthesis and relighting. In this paper, we propose a novel method for representing a scene illuminated by a point light using a set of relightable 3D Gaussian points. Inspired by the Blinn-Phong model, our approach decomposes the scene into ambient, diffuse, and specular components, enabling the synthesis of realistic lighting effects. To facilitate the decomposition of geometric information independent of lighting conditions, we introduce a novel bilevel optimization-based meta-learning framework. The fundamental idea is to view the rendering tasks under various lighting positions as a multi-task learning problem, which our meta-learning approach effectively addresses by generalizing the learned Gaussian geometries not only across different viewpoints but also across diverse light positions. Experimental results demonstrate the effectiveness of our approach in terms of training efficiency and rendering quality compared to existing methods for free-viewpoint relighting.
Authors: Huaxiang Zhang, Yaojia Mu, Guo-Niu Zhu, Zhongxue Gan
Abstract: Accurate visual understanding is imperative for advancing autonomous systems and intelligent robots. Despite the powerful capabilities of vision-language models (VLMs) in processing complex visual scenes, precisely recognizing obscured or ambiguously presented visual elements remains challenging. To tackle such issues, this paper proposes InsightSee, a multi-agent framework to enhance VLMs' interpretative capabilities in handling complex visual understanding scenarios. The framework comprises a description agent, two reasoning agents, and a decision agent, which are integrated to refine the process of visual information interpretation. The design of these agents and the mechanisms by which they can be enhanced in visual information processing are presented. Experimental results demonstrate that the InsightSee framework not only boosts performance on specific visual tasks but also retains the original models' strength. The proposed framework outperforms state-of-the-art algorithms in 6 out of 9 benchmark tests, with a substantial advancement in multimodal understanding.
Authors: Shiyin Lu, Yang Li, Qing-Guo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, Han-Jia Ye
Abstract: Current Multimodal Large Language Models (MLLMs) typically integrate a pre-trained LLM with another pre-trained vision transformer through a connector, such as an MLP, endowing the LLM with visual capabilities. However, the misalignment between two embedding strategies in MLLMs -- the structural textual embeddings based on an embedding look-up table and the continuous embeddings generated directly by the vision encoder -- makes challenges for a more seamless fusion of visual and textual information. We propose Ovis, a novel MLLM architecture designed to structurally align visual and textual embeddings. Ovis integrates an additional learnable visual embedding table into the visual encoder's process. To capture rich visual semantics, each image patch indexes the visual embedding table multiple times, resulting in a final visual embedding that is a probabilistic combination of the indexed embeddings. This structural approach mirrors the method used for generating textual embeddings. Empirical evaluations on various multimodal benchmarks demonstrate that Ovis outperforms open-source MLLMs of similar parameter scales and even surpasses the proprietary model Qwen-VL-Plus overall. These results highlight the potential of Ovis' structured visual representation for advancing MLLM architectural design and promoting more effective multimodal learning. Both the source code and the training dataset of Ovis will be made publicly available.
Authors: Yunbin Tu, Liang Li, Li Su, Zheng-Jun Zha, Chenggang Yan, Qingming Huang
Abstract: Multi-change captioning aims to describe complex and coupled changes within an image pair in natural language. Compared with single-change captioning, this task requires the model to have higher-level cognition ability to reason an arbitrary number of changes. In this paper, we propose a novel context-aware difference distilling (CARD) network to capture all genuine changes for yielding sentences. Given an image pair, CARD first decouples context features that aggregate all similar/dissimilar semantics, termed common/difference context features. Then, the consistency and independence constraints are designed to guarantee the alignment/discrepancy of common/difference context features. Further, the common context features guide the model to mine locally unchanged features, which are subtracted from the pair to distill locally difference features. Next, the difference context features augment the locally difference features to ensure that all changes are distilled. In this way, we obtain an omni-representation of all changes, which is translated into linguistic sentences by a transformer decoder. Extensive experiments on three public datasets show CARD performs favourably against state-of-the-art methods.The code is available at https://github.com/tuyunbin/CARD.
Authors: Seongheon Park, Hyuk Kwon, Kwanghoon Sohn, Kibok Lee
Abstract: Open-world semi-supervised learning (OWSSL) extends conventional semi-supervised learning to open-world scenarios by taking account of novel categories in unlabeled datasets. Despite the recent advancements in OWSSL, the success often relies on the assumptions that 1) labeled and unlabeled datasets share the same balanced class prior distribution, which does not generally hold in real-world applications, and 2) unlabeled training datasets are utilized for evaluation, where such transductive inference might not adequately address challenges in the wild. In this paper, we aim to generalize OWSSL by addressing them. Our work suggests that practical OWSSL may require different training settings, evaluation methods, and learning strategies compared to those prevalent in the existing literature.
Authors: Cheng Tan, Jingxuan Wei, Linzhuang Sun, Zhangyang Gao, Siyuan Li, Bihui Yu, Ruifeng Guo, Stan Z. Li
Abstract: Large language models equipped with retrieval-augmented generation (RAG) represent a burgeoning field aimed at enhancing answering capabilities by leveraging external knowledge bases. Although the application of RAG with language-only models has been extensively explored, its adaptation into multimodal vision-language models remains nascent. Going beyond mere answer generation, the primary goal of multimodal RAG is to cultivate the models' ability to reason in response to relevant queries. To this end, we introduce a novel multimodal RAG framework named RMR (Retrieval Meets Reasoning). The RMR framework employs a bi-modal retrieval module to identify the most relevant question-answer pairs, which then serve as scaffolds for the multimodal reasoning process. This training-free approach not only encourages the model to engage deeply with the reasoning processes inherent in the retrieved content but also facilitates the generation of answers that are precise and richly interpretable. Surprisingly, utilizing solely the ScienceQA dataset, collected from elementary and high school science curricula, RMR significantly boosts the performance of various vision-language models across a spectrum of benchmark datasets, including A-OKVQA, MMBench, and SEED. These outcomes highlight the substantial potential of our multimodal retrieval and reasoning mechanism to improve the reasoning capabilities of vision-language models.
Authors: Shurong Yang, Huadong Li, Juhao Wu, Minhao Jing, Linze Li, Renhe Ji, Jiajun Liang, Haoqiang Fan
Abstract: Despite raw driving videos contain richer information on facial expressions than intermediate representations such as landmarks in the field of portrait animation, they are seldom the subject of research. This is due to two challenges inherent in portrait animation driven with raw videos: 1) significant identity leakage; 2) Irrelevant background and facial details such as wrinkles degrade performance. To harnesses the power of the raw videos for vivid portrait animation, we proposed a pioneering conditional diffusion model named as MegActor. First, we introduced a synthetic data generation framework for creating videos with consistent motion and expressions but inconsistent IDs to mitigate the issue of ID leakage. Second, we segmented the foreground and background of the reference image and employed CLIP to encode the background details. This encoded information is then integrated into the network via a text embedding module, thereby ensuring the stability of the background. Finally, we further style transfer the appearance of the reference image to the driving video to eliminate the influence of facial details in the driving videos. Our final model was trained solely on public datasets, achieving results comparable to commercial models. We hope this will help the open-source community.The code is available at https://github.com/megvii-research/MegFaceAnimate.
Authors: Sijin Chen, Xin Chen, Anqi Pang, Xianfang Zeng, Wei Cheng, Yijun Fu, Fukun Yin, Yanru Wang, Zhibin Wang, Chi Zhang, Jingyi Yu, Gang Yu, Bin Fu, Tao Chen
Abstract: The polygon mesh representation of 3D data exhibits great flexibility, fast rendering speed, and storage efficiency, which is widely preferred in various applications. However, given its unstructured graph representation, the direct generation of high-fidelity 3D meshes is challenging. Fortunately, with a pre-defined ordering strategy, 3D meshes can be represented as sequences, and the generation process can be seamlessly treated as an auto-regressive problem. In this paper, we validate the Neural Coordinate Field (NeurCF), an explicit coordinate representation with implicit neural embeddings, is a simple-yet-effective representation for large-scale sequential mesh modeling. After that, we present MeshXL, a family of generative pre-trained auto-regressive models, which addresses the process of 3D mesh generation with modern large language model approaches. Extensive experiments show that MeshXL is able to generate high-quality 3D meshes, and can also serve as foundation models for various down-stream applications.
Authors: Eunho Lee, Youngbae Hwang
Abstract: Despite the strong performance of Transformers, their quadratic computation complexity presents challenges in applying them to vision tasks. Automatic pruning is one of effective methods for reducing computation complexity without heuristic approaches. However, directly applying it to multi-head attention is not straightforward due to channel misalignment. In this paper, we propose an automatic channel pruning method to take into account the multi-head attention mechanism. First, we incorporate channel similarity-based weights into the pruning indicator to preserve more informative channels in each head. Then, we adjust pruning indicator to enforce removal of channels in equal proportions across all heads, preventing the channel misalignment. We also add a reweight module to compensate for information loss resulting from channel removal, and an effective initialization step for pruning indicator based on difference of attention between original structure and each channel. Our proposed method can be used to not only original attention, but also linear attention, which is more efficient as linear complexity with respect to the number of tokens. On ImageNet-1K, applying our pruning method to the FLattenTransformer, which includes both attention mechanisms, shows outperformed accuracy for several MACs compared with previous state-of-the-art efficient models and pruned methods. Code will be available soon.
Authors: Pedram Ghamisi, Weikang Yu, Andrea Marinoni, Caroline M. Gevaert, Claudio Persello, Sivasakthy Selvakumaran, Manuela Girotto, Benjamin P. Horton, Philippe Rufin, Patrick Hostert, Fabio Pacifici, Peter M. Atkinson
Abstract: The convergence of artificial intelligence (AI) and Earth observation (EO) technologies has brought geoscience and remote sensing into an era of unparalleled capabilities. AI's transformative impact on data analysis, particularly derived from EO platforms, holds great promise in addressing global challenges such as environmental monitoring, disaster response and climate change analysis. However, the rapid integration of AI necessitates a careful examination of the responsible dimensions inherent in its application within these domains. In this paper, we represent a pioneering effort to systematically define the intersection of AI and EO, with a central focus on responsible AI practices. Specifically, we identify several critical components guiding this exploration from both academia and industry perspectives within the EO field: AI and EO for social good, mitigating unfair biases, AI security in EO, geo-privacy and privacy-preserving measures, as well as maintaining scientific excellence, open data, and guiding AI usage based on ethical principles. Furthermore, the paper explores potential opportunities and emerging trends, providing valuable insights for future research endeavors.
Authors: Pallavi Mitra, Gesina Schwalbe, Nadja Klein
Abstract: Convolutional Neural Networks (CNNs) have achieved state-of-the-art performance in many computer vision tasks. However, high computational and storage demands hinder their deployment into resource-constrained environments, such as embedded devices. Model pruning helps to meet these restrictions by reducing the model size, while maintaining superior performance. Meanwhile, safety-critical applications pose more than just resource and performance constraints. In particular, predictions must not be overly confident, i.e., provide properly calibrated uncertainty estimations (proper uncertainty calibration), and CNNs must be robust against corruptions like naturally occurring input perturbations (natural corruption robustness). This work investigates the important trade-off between uncertainty calibration, natural corruption robustness, and performance for current state-of-research post-hoc CNN pruning techniques in the context of image classification tasks. Our study reveals that post-hoc pruning substantially improves the model's uncertainty calibration, performance, and natural corruption robustness, sparking hope for safe and robust embedded CNNs.Furthermore, uncertainty calibration and natural corruption robustness are not mutually exclusive targets under pruning, as evidenced by the improved safety aspects obtained by post-hoc unstructured pruning with increasing compression.
Authors: Haolong Ma, Hui Li, Chunyang Cheng, Gaoang Wang, Xiaoning Song, Xiaojun Wu
Abstract: As one of the tasks in Image Fusion, Infrared and Visible Image Fusion aims to integrate complementary information captured by sensors of different modalities into a single image. The Selective State Space Model (SSSM), known for its ability to capture long-range dependencies, has demonstrated its potential in the field of computer vision. However, in image fusion, current methods underestimate the potential of SSSM in capturing the global spatial information of both modalities. This limitation prevents the simultaneous consideration of the global spatial information from both modalities during interaction, leading to a lack of comprehensive perception of salient targets. Consequently, the fusion results tend to bias towards one modality instead of adaptively preserving salient targets. To address this issue, we propose the Saliency-aware Selective State Space Fusion Model (S4Fusion). In our S4Fusion, the designed Cross-Modal Spatial Awareness Module (CMSA) can simultaneously focus on global spatial information from both modalities while facilitating their interaction, thereby comprehensively capturing complementary information. Additionally, S4Fusion leverages a pre-trained network to perceive uncertainty in the fused images. By minimizing this uncertainty, S4Fusion adaptively highlights salient targets from both images. Extensive experiments demonstrate that our approach produces high-quality images and enhances performance in downstream tasks.
Authors: Zhipeng Yang, Ruoyu Wang, Yang Tan, Liping Xie
Abstract: Online action detection (OAD) aims to identify ongoing actions from streaming video in real-time, without access to future frames. Since these actions manifest at varying scales of granularity, ranging from coarse to fine, projecting an entire set of action frames to a single latent encoding may result in a lack of local information, necessitating the acquisition of action features across multiple scales. In this paper, we propose a multi-scale action learning transformer (MALT), which includes a novel recurrent decoder (used for feature fusion) that includes fewer parameters and can be trained more efficiently. A hierarchical encoder with multiple encoding branches is further proposed to capture multi-scale action features. The output from the preceding branch is then incrementally input to the subsequent branch as part of a cross-attention calculation. In this way, output features transition from coarse to fine as the branches deepen. We also introduce an explicit frame scoring mechanism employing sparse attention, which filters irrelevant frames more efficiently, without requiring an additional network. The proposed method achieved state-of-the-art performance on two benchmark datasets (THUMOS'14 and TVSeries), outperforming all existing models used for comparison, with an mAP of 0.2% for THUMOS'14 and an mcAP of 0.1% for TVseries.
Authors: Adithya TG, Adithya SK, Abhinav R Bharadwaj, Abhiram HA, Dr. Surabhi Narayan
Abstract: Interacting and understanding with text heavy visual content with multiple images is a major challenge for traditional vision models. This paper is on enhancing vision models' capability to comprehend or understand and learn from images containing a huge amount of textual information from the likes of textbooks and research papers which contain multiple images like graphs, etc and tables in them with different types of axes and scales. The approach involves dataset preprocessing, fine tuning which is by using instructional oriented data and evaluation. We also built a visual chat application integrating CLIP for image encoding and a model from the Massive Text Embedding Benchmark which is developed to consider both textual and visual inputs. An accuracy of 96.71% was obtained. The aim of the project is to increase and also enhance the advance vision models' capabilities in understanding complex visual textual data interconnected data, contributing to multimodal AI.
Authors: Felix Mujkanovic, Ntumba Elie Nsampi, Christian Theobalt, Hans-Peter Seidel, Thomas Leimk\"uhler
Abstract: Gaussian scale spaces are a cornerstone of signal representation and processing, with applications in filtering, multiscale analysis, anti-aliasing, and many more. However, obtaining such a scale space is costly and cumbersome, in particular for continuous representations such as neural fields. We present an efficient and lightweight method to learn the fully continuous, anisotropic Gaussian scale space of an arbitrary signal. Based on Fourier feature modulation and Lipschitz bounding, our approach is trained self-supervised, i.e., training does not require any manual filtering. Our neural Gaussian scale-space fields faithfully capture multiscale representations across a broad range of modalities, and support a diverse set of applications. These include images, geometry, light-stage data, texture anti-aliasing, and multiscale optimization.
Authors: Linli Yao, Lei Li, Shuhuai Ren, Lean Wang, Yuanxin Liu, Xu Sun, Lu Hou
Abstract: The visual projector, which bridges the vision and language modalities and facilitates cross-modal alignment, serves as a crucial component in MLLMs. However, measuring the effectiveness of projectors in vision-language alignment remains under-explored, which currently can only be inferred from the performance of MLLMs on downstream tasks. Motivated by the problem, this study examines the projector module by interpreting the vision-language semantic flow within MLLMs. Specifically, we trace back the semantic relevance flow from generated language tokens to raw visual encoder patches and the intermediate outputs produced by projectors. Our findings reveal that compressive projectors (e.g., QFormer), abstract visual patches into a limited set of semantic concepts, such as objects or attributes, resulting in a 'double abstraction' phenomenon. This involves a first visual semantic abstraction by the projector referring to pre-defined query tokens, and a second extraction by the LLM based on text instructions. The double abstraction is inefficient in training and will result in cumulative vision semantics deficiency. To mitigate this issue, we propose the key insight of 'Decouple Compression from Abstraction (DeCo), that is compressing the visual token number at the patch level by projectors and allowing the LLM to handle visual semantic abstraction entirely. Consequently, we adopt a simple compressor, i.e., 2D Adaptive Pooling, to downsample visual patches in a parameter-free manner. Empirical evaluation demonstrates that DeCo surpasses traditional compressive projectors regarding both performance and efficiency. It achieves performance gains of 0.9%, 7.1%, and 2.9% across the MLLM Benchmarks, Visual Localization, and Open-ended VQA tasks with fewer trainable parameters and faster convergence speed.
Authors: Muhammad Muneeb Saad, Mubashir Husain Rehmani, Ruairi O'Reilly
Abstract: Generative Adversarial Networks (GANs) have high computational costs to train their complex architectures. Throughout the training process, GANs' output is analyzed qualitatively based on the loss and synthetic images' diversity and quality. Based on this qualitative analysis, training is manually halted once the desired synthetic images are generated. By utilizing an early stopping criterion, the computational cost and dependence on manual oversight can be reduced yet impacted by training problems such as mode collapse, non-convergence, and instability. This is particularly prevalent in biomedical imagery, where training problems degrade the diversity and quality of synthetic images, and the high computational cost associated with training makes complex architectures increasingly inaccessible. This work proposes a novel early stopping criteria to quantitatively detect training problems, halt training, and reduce the computational costs associated with synthesizing biomedical images. Firstly, the range of generator and discriminator loss values is investigated to assess whether mode collapse, non-convergence, and instability occur sequentially, concurrently, or interchangeably throughout the training of GANs. Secondly, utilizing these occurrences in conjunction with the Mean Structural Similarity Index (MS-SSIM) and Fr\'echet Inception Distance (FID) scores of synthetic images forms the basis of the proposed early stopping criteria. This work helps identify the occurrence of training problems in GANs using low-resource computational cost and reduces training time to generate diversified and high-quality synthetic images.
Authors: Yi Yang, Qingwen Zhang, Kei Ikemura, Nazre Batool, John Folkesson
Abstract: Addressing hard cases in autonomous driving, such as anomalous road users, extreme weather conditions, and complex traffic interactions, presents significant challenges. To ensure safety, it is crucial to detect and manage these scenarios effectively for autonomous driving systems. However, the rarity and high-risk nature of these cases demand extensive, diverse datasets for training robust models. Vision-Language Foundation Models (VLMs) have shown remarkable zero-shot capabilities as being trained on extensive datasets. This work explores the potential of VLMs in detecting hard cases in autonomous driving. We demonstrate the capability of VLMs such as GPT-4v in detecting hard cases in traffic participant motion prediction on both agent and scenario levels. We introduce a feasible pipeline where VLMs, fed with sequential image frames with designed prompts, effectively identify challenging agents or scenarios, which are verified by existing prediction models. Moreover, by taking advantage of this detection of hard cases by VLMs, we further improve the training efficiency of the existing motion prediction pipeline by performing data selection for the training samples suggested by GPT. We show the effectiveness and feasibility of our pipeline incorporating VLMs with state-of-the-art methods on NuScenes datasets. The code is accessible at https://github.com/KTH-RPL/Detect_VLM.
Authors: Pengyuan Lyu, Yulin Li, Hao Zhou, Weihong Ma, Xingyu Wan, Qunyi Xie, Liang Wu, Chengquan Zhang, Kun Yao, Errui Ding, Jingdong Wang
Abstract: Text-rich images have significant and extensive value, deeply integrated into various aspects of human life. Notably, both visual cues and linguistic symbols in text-rich images play crucial roles in information transmission but are accompanied by diverse challenges. Therefore, the efficient and effective understanding of text-rich images is a crucial litmus test for the capability of Vision-Language Models. We have crafted an efficient vision-language model, StrucTexTv3, tailored to tackle various intelligent tasks for text-rich images. The significant design of StrucTexTv3 is presented in the following aspects: Firstly, we adopt a combination of an effective multi-scale reduced visual transformer and a multi-granularity token sampler (MG-Sampler) as a visual token generator, successfully solving the challenges of high-resolution input and complex representation learning for text-rich images. Secondly, we enhance the perception and comprehension abilities of StrucTexTv3 through instruction learning, seamlessly integrating various text-oriented tasks into a unified framework. Thirdly, we have curated a comprehensive collection of high-quality text-rich images, abbreviated as TIM-30M, encompassing diverse scenarios like incidental scenes, office documents, web pages, and screenshots, thereby improving the robustness of our model. Our method achieved SOTA results in text-rich image perception tasks, and significantly improved performance in comprehension tasks. Among multimodal models with LLM decoder of approximately 1.8B parameters, it stands out as a leader, which also makes the deployment of edge devices feasible. In summary, the StrucTexTv3 model, featuring efficient structural design, outstanding performance, and broad adaptability, offers robust support for diverse intelligent application tasks involving text-rich images, thus exhibiting immense potential for widespread application.
Authors: Fatema Jannat Dihan, Saydul Akbar Murad, Abu Jafar Md Muzahid, K. M. Aslam Uddin, Mohammed J. F. Alenazi, Anupam Kumar Bairagi, Sujit Biswas
Abstract: Monkeypox virus (MPXV) is a zoonotic virus that poses a significant threat to public health, particularly in remote parts of Central and West Africa. Early detection of monkeypox lesions is crucial for effective treatment. However, due to its similarity with other skin diseases, monkeypox lesion detection is a challenging task. To detect monkeypox, many researchers used various deep-learning models such as MobileNetv2, VGG16, ResNet50, InceptionV3, DenseNet121, EfficientNetB3, MobileNetV2, and Xception. However, these models often require high storage space due to their large size. This study aims to improve the existing challenges by introducing a CNN model named MpoxSLDNet (Monkeypox Skin Lesion Detector Network) to facilitate early detection and categorization of Monkeypox lesions and Non-Monkeypox lesions in digital images. Our model represents a significant advancement in the field of monkeypox lesion detection by offering superior performance metrics, including precision, recall, F1-score, accuracy, and AUC, compared to traditional pre-trained models such as VGG16, ResNet50, and DenseNet121. The key novelty of our approach lies in MpoxSLDNet's ability to achieve high detection accuracy while requiring significantly less storage space than existing models. By addressing the challenge of high storage requirements, MpoxSLDNet presents a practical solution for early detection and categorization of monkeypox lesions in resource-constrained healthcare settings. In this study, we have used "Monkeypox Skin Lesion Dataset" comprising 1428 skin images of monkeypox lesions and 1764 skin images of Non-Monkeypox lesions. Dataset's limitations could potentially impact the model's ability to generalize to unseen cases. However, the MpoxSLDNet model achieved a validation accuracy of 94.56%, compared to 86.25%, 84.38%, and 67.19% for VGG16, DenseNet121, and ResNet50, respectively.
Authors: Jiatao Gu, Ying Shen, Shuangfei Zhai, Yizhe Zhang, Navdeep Jaitly, Joshua M. Susskind
Abstract: Diffusion models have emerged as a powerful tool for generating high-quality images from textual descriptions. Despite their successes, these models often exhibit limited diversity in the sampled images, particularly when sampling with a high classifier-free guidance weight. To address this issue, we present Kaleido, a novel approach that enhances the diversity of samples by incorporating autoregressive latent priors. Kaleido integrates an autoregressive language model that encodes the original caption and generates latent variables, serving as abstract and intermediary representations for guiding and facilitating the image generation process. In this paper, we explore a variety of discrete latent representations, including textual descriptions, detection bounding boxes, object blobs, and visual tokens. These representations diversify and enrich the input conditions to the diffusion models, enabling more diverse outputs. Our experimental results demonstrate that Kaleido effectively broadens the diversity of the generated image samples from a given textual description while maintaining high image quality. Furthermore, we show that Kaleido adheres closely to the guidance provided by the generated latent variables, demonstrating its capability to effectively control and direct the image generation process.
Authors: Xinxi Zhang, Song Wen, Ligong Han, Felix Juefei-Xu, Akash Srivastava, Junzhou Huang, Hao Wang, Molei Tao, Dimitris N. Metaxas
Abstract: Adapting large-scale pre-trained generative models in a parameter-efficient manner is gaining traction. Traditional methods like low rank adaptation achieve parameter efficiency by imposing constraints but may not be optimal for tasks requiring high representation capacity. We propose a novel spectrum-aware adaptation framework for generative models. Our method adjusts both singular values and their basis vectors of pretrained weights. Using the Kronecker product and efficient Stiefel optimizers, we achieve parameter-efficient adaptation of orthogonal matrices. We introduce Spectral Orthogonal Decomposition Adaptation (SODA), which balances computational efficiency and representation capacity. Extensive evaluations on text-to-image diffusion models demonstrate SODA's effectiveness, offering a spectrum-aware alternative to existing fine-tuning methods.
Authors: Jingjing Wang, Dan Zhang, Feng Luo
Abstract: Previous work has demonstrated that, in the Variance Preserving (VP) scenario, the nascent Directly Denoising Diffusion Models (DDDM) can generate high-quality images in one step while achieving even better performance in multistep sampling. However, the Pseudo-LPIPS loss used in DDDM leads to concerns about the bias in assessment. Here, we propose a unified DDDM (uDDDM) framework that generates images in one-step/multiple steps for both Variance Preserving (VP) and Variance Exploding (VE) cases. We provide theoretical proofs of the existence and uniqueness of the model's solution paths, as well as the non-intersecting property of the sampling paths. Additionally, we propose an adaptive Pseudo-Huber loss function to balance the convergence to the true solution and the stability of convergence process.Through a comprehensive evaluation, we demonstrate that uDDDMs achieve FID scores comparable to the best-performing methods available for CIFAR-10 in both VP and VE. Specifically, uDDDM achieves one-step generation on CIFAR10 with FID of 2.63 and 2.53 for VE and VP respectively. By extending the sampling to 1000 steps, we further reduce FID score to 1.71 and 1.65 for VE and VP respectively, setting state-of-the-art performance in both cases.
Authors: Siyi Hu, Diego Martin Arroyo, Stephanie Debats, Fabian Manhardt, Luca Carlone, Federico Tombari
Abstract: Realistic conditional 3D scene synthesis significantly enhances and accelerates the creation of virtual environments, which can also provide extensive training data for computer vision and robotics research among other applications. Diffusion models have shown great performance in related applications, e.g., making precise arrangements of unordered sets. However, these models have not been fully explored in floor-conditioned scene synthesis problems. We present MiDiffusion, a novel mixed discrete-continuous diffusion model architecture, designed to synthesize plausible 3D indoor scenes from given room types, floor plans, and potentially pre-existing objects. We represent a scene layout by a 2D floor plan and a set of objects, each defined by its category, location, size, and orientation. Our approach uniquely implements structured corruption across the mixed discrete semantic and continuous geometric domains, resulting in a better conditioned problem for the reverse denoising step. We evaluate our approach on the 3D-FRONT dataset. Our experimental results demonstrate that MiDiffusion substantially outperforms state-of-the-art autoregressive and diffusion models in floor-conditioned 3D scene synthesis. In addition, our models can handle partial object constraints via a corruption-and-masking strategy without task specific training. We show MiDiffusion maintains clear advantages over existing approaches in scene completion and furniture arrangement experiments.
Authors: Xin Wen, Bingchen Zhao, Yilun Chen, Jiangmiao Pang, Xiaojuan Qi
Abstract: Severe data imbalance naturally exists among web-scale vision-language datasets. Despite this, we find CLIP pre-trained thereupon exhibits notable robustness to the data imbalance compared to supervised learning, and demonstrates significant effectiveness in learning generalizable representations. With an aim to investigate the reasons behind this finding, we conduct controlled experiments to study various underlying factors, and reveal that CLIP's pretext task forms a dynamic classification problem wherein only a subset of classes is present in training. This isolates the bias from dominant classes and implicitly balances the learning signal. Furthermore, the robustness and discriminability of CLIP improve with more descriptive language supervision, larger data scale, and broader open-world concepts, which are inaccessible to supervised learning. Our study not only uncovers the mechanisms behind CLIP's generalizability beyond data imbalance but also provides transferable insights for the research community. The findings are validated in both supervised and self-supervised learning, enabling models trained on imbalanced data to achieve CLIP-level performance on diverse recognition tasks. Code will be available at: https://github.com/CVMI-Lab/clip-beyond-tail.
Authors: Xiao Zhang, William Gao, Seemandhar Jain, Michael Maire, David. A. Forsyth, Anand Bhattad
Abstract: Image relighting is the task of showing what a scene from a source image would look like if illuminated differently. Inverse graphics schemes recover an explicit representation of geometry and a set of chosen intrinsics, then relight with some form of renderer. However error control for inverse graphics is difficult, and inverse graphics methods can represent only the effects of the chosen intrinsics. This paper describes a relighting method that is entirely data-driven, where intrinsics and lighting are each represented as latent variables. Our approach produces SOTA relightings of real scenes, as measured by standard metrics. We show that albedo can be recovered from our latent intrinsics without using any example albedos, and that the albedos recovered are competitive with SOTA methods.
Authors: Chaoyou Fu, Yuhan Dai, Yondong Luo, Lei Li, Shuhuai Ren, Renrui Zhang, Zihan Wang, Chenyu Zhou, Yunhang Shen, Mengdan Zhang, Peixian Chen, Yanwei Li, Shaohui Lin, Sirui Zhao, Ke Li, Tong Xu, Xiawu Zheng, Enhong Chen, Rongrong Ji, Xing Sun
Abstract: In the quest for artificial general intelligence, Multi-modal Large Language Models (MLLMs) have emerged as a focal point in recent advancements. However, the predominant focus remains on developing their capabilities in static image understanding. The potential of MLLMs in processing sequential visual data is still insufficiently explored, highlighting the absence of a comprehensive, high-quality assessment of their performance. In this paper, we introduce Video-MME, the first-ever full-spectrum, Multi-Modal Evaluation benchmark of MLLMs in Video analysis. Our work distinguishes from existing benchmarks through four key features: 1) Diversity in video types, spanning 6 primary visual domains with 30 subfields to ensure broad scenario generalizability; 2) Duration in temporal dimension, encompassing both short-, medium-, and long-term videos, ranging from 11 seconds to 1 hour, for robust contextual dynamics; 3) Breadth in data modalities, integrating multi-modal inputs besides video frames, including subtitles and audios, to unveil the all-round capabilities of MLLMs; 4) Quality in annotations, utilizing rigorous manual labeling by expert annotators to facilitate precise and reliable model assessment. 900 videos with a total of 256 hours are manually selected and annotated by repeatedly viewing all the video content, resulting in 2,700 question-answer pairs. With Video-MME, we extensively evaluate various state-of-the-art MLLMs, including GPT-4 series and Gemini 1.5 Pro, as well as open-source image models like InternVL-Chat-V1.5 and video models like LLaVA-NeXT-Video. Our experiments reveal that Gemini 1.5 Pro is the best-performing commercial model, significantly outperforming the open-source models. Our dataset along with these findings underscores the need for further improvements in handling longer sequences and multi-modal data. Project Page: https://video-mme.github.io
Authors: Yujia Liu, Tong Bu, Jianhao Ding, Zecheng Hao, Tiejun Huang, Zhaofei Yu
Abstract: Spiking Neural Networks (SNNs) have attracted great attention for their energy-efficient operations and biologically inspired structures, offering potential advantages over Artificial Neural Networks (ANNs) in terms of energy efficiency and interpretability. Nonetheless, similar to ANNs, the robustness of SNNs remains a challenge, especially when facing adversarial attacks. Existing techniques, whether adapted from ANNs or specifically designed for SNNs, exhibit limitations in training SNNs or defending against strong attacks. In this paper, we propose a novel approach to enhance the robustness of SNNs through gradient sparsity regularization. We observe that SNNs exhibit greater resilience to random perturbations compared to adversarial perturbations, even at larger scales. Motivated by this, we aim to narrow the gap between SNNs under adversarial and random perturbations, thereby improving their overall robustness. To achieve this, we theoretically prove that this performance gap is upper bounded by the gradient sparsity of the probability associated with the true label concerning the input image, laying the groundwork for a practical strategy to train robust SNNs by regularizing the gradient sparsity. We validate the effectiveness of our approach through extensive experiments on both image-based and event-based datasets. The results demonstrate notable improvements in the robustness of SNNs. Our work highlights the importance of gradient sparsity in SNNs and its role in enhancing robustness.
Authors: Jiyue Huang, Chi Hong, Lydia Y. Chen, Stefanie Roos
Abstract: Diffusion models are becoming defector generative models, which generate exceptionally high-resolution image data. Training effective diffusion models require massive real data, which is privately owned by distributed parties. Each data party can collaboratively train diffusion models in a federated learning manner by sharing gradients instead of the raw data. In this paper, we study the privacy leakage risk of gradient inversion attacks. First, we design a two-phase fusion optimization, GIDM, to leverage the well-trained generative model itself as prior knowledge to constrain the inversion search (latent) space, followed by pixel-wise fine-tuning. GIDM is shown to be able to reconstruct images almost identical to the original ones. Considering a more privacy-preserving training scenario, we then argue that locally initialized private training noise $\epsilon$ and sampling step t may raise additional challenges for the inversion attack. To solve this, we propose a triple-optimization GIDM+ that coordinates the optimization of the unknown data, $\epsilon$ and $t$. Our extensive evaluation results demonstrate the vulnerability of sharing gradient for data protection of diffusion models, even high-resolution images can be reconstructed with high quality.
Authors: Egor Kashkarov, Egor Chistov, Ivan Molodetskikh, Dmitriy Vatolin
Abstract: Perceptual losses play an important role in constructing deep-neural-network-based methods by increasing the naturalness and realism of processed images and videos. Use of perceptual losses is often limited to LPIPS, a fullreference method. Even though deep no-reference image-qualityassessment methods are excellent at predicting human judgment, little research has examined their incorporation in loss functions. This paper investigates direct optimization of several video-superresolution models using no-reference image-quality-assessment methods as perceptual losses. Our experimental results show that straightforward optimization of these methods produce artifacts, but a special training procedure can mitigate them.
Authors: Haibo Jin, Andy Zhou, Joe D. Menke, Haohan Wang
Abstract: Large Language Models (LLMs) are typically harmless but remain vulnerable to carefully crafted prompts known as ``jailbreaks'', which can bypass protective measures and induce harmful behavior. Recent advancements in LLMs have incorporated moderation guardrails that can filter outputs, which trigger processing errors for certain malicious questions. Existing red-teaming benchmarks often neglect to include questions that trigger moderation guardrails, making it difficult to evaluate jailbreak effectiveness. To address this issue, we introduce JAMBench, a harmful behavior benchmark designed to trigger and evaluate moderation guardrails. JAMBench involves 160 manually crafted instructions covering four major risk categories at multiple severity levels. Furthermore, we propose a jailbreak method, JAM (Jailbreak Against Moderation), designed to attack moderation guardrails using jailbreak prefixes to bypass input-level filters and a fine-tuned shadow model functionally equivalent to the guardrail model to generate cipher characters to bypass output-level filters. Our extensive experiments on four LLMs demonstrate that JAM achieves higher jailbreak success ($\sim$ $\times$ 19.88) and lower filtered-out rates ($\sim$ $\times$ 1/6) than baselines.
Authors: Levy Chaves, Eduardo Valle, Alceu Bissoto, Sandra Avila
Abstract: In the evolving landscape of deep learning, selecting the best pre-trained models from a growing number of choices is a challenge. Transferability scorers propose alleviating this scenario, but their recent proliferation, ironically, poses the challenge of their own assessment. In this work, we propose both robust benchmark guidelines for transferability scorers, and a well-founded technique to combine multiple scorers, which we show consistently improves their results. We extensively evaluate 13 scorers from literature across 11 datasets, comprising generalist, fine-grained, and medical imaging datasets. We show that few scorers match the predictive performance of the simple raw metric of models on ImageNet, and that all predictors suffer on medical datasets. Our results highlight the potential of combining different information sources for reliably predicting transferability across varied domains.
Authors: Khiem Le, Nhan Luong-Ha, Manh Nguyen-Duc, Danh Le-Phuoc, Cuong Do, Kok-Seng Wong
Abstract: Federated Learning (FL) is a promising paradigm that offers significant advancements in privacy-preserving, decentralized machine learning by enabling collaborative training of models across distributed devices without centralizing data. However, the practical deployment of FL systems faces a significant bottleneck: the communication overhead caused by frequently exchanging large model updates between numerous devices and a central server. This communication inefficiency can hinder training speed, model performance, and the overall feasibility of real-world FL applications. In this survey, we investigate various strategies and advancements made in communication-efficient FL, highlighting their impact and potential to overcome the communication challenges inherent in FL systems. Specifically, we define measures for communication efficiency, analyze sources of communication inefficiency in FL systems, and provide a taxonomy and comprehensive review of state-of-the-art communication-efficient FL methods. Additionally, we discuss promising future research directions for enhancing the communication efficiency of FL systems. By addressing the communication bottleneck, FL can be effectively applied and enable scalable and practical deployment across diverse applications that require privacy-preserving, decentralized machine learning, such as IoT, healthcare, or finance.
Authors: Jiuhong Xiao, Ning Zhang, Daniel Tortei, Giuseppe Loianno
Abstract: Accurate geo-localization of Unmanned Aerial Vehicles (UAVs) is crucial for a variety of outdoor applications including search and rescue operations, power line inspections, and environmental monitoring. The vulnerability of Global Navigation Satellite Systems (GNSS) signals to interference and spoofing necessitates the development of additional robust localization methods for autonomous navigation. Visual Geo-localization (VG), leveraging onboard cameras and reference satellite maps, offers a promising solution for absolute localization. Specifically, Thermal Geo-localization (TG), which relies on image-based matching between thermal imagery with satellite databases, stands out by utilizing infrared cameras for effective night-time localization. However, the efficiency and effectiveness of current TG approaches, are hindered by dense sampling on satellite maps and geometric noises in thermal query images. To overcome these challenges, in this paper, we introduce STHN, a novel UAV thermal geo-localization approach that employs a coarse-to-fine deep homography estimation method. This method attains reliable thermal geo-localization within a 512-meter radius of the UAV's last known location even with a challenging 11% overlap between satellite and thermal images, despite the presence of indistinct textures in thermal imagery and self-similar patterns in both spectra. Our research significantly enhances UAV thermal geo-localization performance and robustness against the impacts of geometric noises under low-visibility conditions in the wild. The code will be made publicly available.
Authors: Shivendra Agrawal, Suresh Nayak, Ashutosh Naik, Bradley Hayes
Abstract: The ability to shop independently, especially in grocery stores, is important for maintaining a high quality of life. This can be particularly challenging for people with visual impairments (PVI). Stores carry thousands of products, with approximately 30,000 new products introduced each year in the US market alone, presenting a challenge even for modern computer vision solutions. Through this work, we present a proof-of-concept socially assistive robotic system we call ShelfHelp, and propose novel technical solutions for enhancing instrumented canes traditionally meant for navigation tasks with additional capability within the domain of shopping. ShelfHelp includes a novel visual product locator algorithm designed for use in grocery stores and a novel planner that autonomously issues verbal manipulation guidance commands to guide the user during product retrieval. Through a human subjects study, we show the system's success in locating and providing effective manipulation guidance to retrieve desired products with novice users. We compare two autonomous verbal guidance modes achieving comparable performance to a human assistance baseline and present encouraging findings that validate our system's efficiency and effectiveness and through positive subjective metrics including competence, intelligence, and ease of use.
Authors: Aastha Acharya, Caleb Lee, Marissa D'Alonzo, Jared Shamwell, Nisar R. Ahmed, Rebecca Russell
Abstract: Deep learning offers promising new ways to accurately model aleatoric uncertainty in robotic estimation systems, particularly when the uncertainty distributions do not conform to traditional assumptions of being fixed and Gaussian. In this study, we formulate and evaluate three fundamental deep learning approaches for conditional probability density modeling to quantify non-Gaussian aleatoric uncertainty: parametric, discretized, and generative modeling. We systematically compare the respective strengths and weaknesses of these three methods on simulated non-Gaussian densities as well as on real-world terrain-relative navigation data. Our results show that these deep learning methods can accurately capture complex uncertainty patterns, highlighting their potential for improving the reliability and robustness of estimation systems.
Authors: Kyle Henke, Elijah Pelofske, Garrett Kenyon, Georg Hahn
Abstract: We consider the problem of computing a sparse binary representation of an image. To be precise, given an image and an overcomplete, non-orthonormal basis, we aim to find a sparse binary vector indicating the minimal set of basis vectors that when added together best reconstruct the given input. We formulate this problem with an $L_2$ loss on the reconstruction error, and an $L_0$ (or, equivalently, an $L_1$) loss on the binary vector enforcing sparsity. This yields a quadratic binary optimization problem (QUBO), whose optimal solution(s) in general is NP-hard to find. The method of unsupervised and unnormalized dictionary feature learning for a desired sparsity level to best match the data is presented. Next, we solve the sparse representation QUBO by implementing it both on a D-Wave quantum annealer with Pegasus chip connectivity via minor embedding, as well as on the Intel Loihi 2 spiking neuromorphic processor. On the quantum annealer, we sample from the sparse representation QUBO using parallel quantum annealing combined with quantum evolution Monte Carlo, also known as iterated reverse annealing. On Loihi 2, we use a stochastic winner take all network of neurons. The solutions are benchmarked against simulated annealing, a classical heuristic, and the optimal solutions are computed using CPLEX. Iterated reverse quantum annealing performs similarly to simulated annealing, although simulated annealing is always able to sample the optimal solution whereas quantum annealing was not always able to. The Loihi 2 solutions that are sampled are on average more sparse than the solutions from any of the other methods. Loihi 2 outperforms a D-Wave quantum annealer standard linear-schedule anneal, while iterated reverse quantum annealing performs much better than both unmodified linear-schedule quantum annealing and iterated warm starting on Loihi 2.
Authors: Henry Pinkard, Leyla Kabuli, Eric Markley, Tiffany Chien, Jiantao Jiao, Laura Waller
Abstract: Information theory, which describes the transmission of signals in the presence of noise, has enabled the development of reliable communication systems that underlie the modern world. Imaging systems can also be viewed as a form of communication, in which information about the object is "transmitted" through images. However, the application of information theory to imaging systems has been limited by the challenges of accounting for their physical constraints. Here, we introduce a framework that addresses these limitations by modeling the probabilistic relationship between objects and their measurements. Using this framework, we develop a method to estimate information using only a dataset of noisy measurements, without making any assumptions about the image formation process. We demonstrate that these estimates comprehensively quantify measurement quality across a diverse range of imaging systems and applications. Furthermore, we introduce Information-Driven Encoder Analysis Learning (IDEAL), a technique to optimize the design of imaging hardware for maximum information capture. This work provides new insights into the fundamental performance limits of imaging systems and offers powerful new tools for their analysis and design.
Authors: Jung H. Lee, Sujith Vijayan
Abstract: Deep learning (DL) enables deep neural networks (DNNs) to automatically learn complex tasks or rules from given examples without instructions or guiding principles. As we do not engineer DNNs' functions, it is extremely difficult to diagnose their decisions, and multiple lines of studies proposed to explain principles of DNNs/DL operations. Notably, one line of studies suggests that DNNs may learn concepts, the high level features recognizable to humans. Thus, we hypothesized that DNNs develop abstract codes, not necessarily recognizable to humans, which can be used to augment DNNs' decision-making. To address this hypothesis, we combined foundation segmentation models and unsupervised learning to extract internal codes and identify potential use of abstract codes to make DL's decision-making more reliable and safer.
Authors: Krishanu Maity, A. S. Poornash, Sriparna Saha, Pushpak Bhattacharyya
Abstract: In an era of rapidly evolving internet technology, the surge in multimodal content, including videos, has expanded the horizons of online communication. However, the detection of toxic content in this diverse landscape, particularly in low-resource code-mixed languages, remains a critical challenge. While substantial research has addressed toxic content detection in textual data, the realm of video content, especially in non-English languages, has been relatively underexplored. This paper addresses this research gap by introducing a benchmark dataset, the first of its kind, consisting of 931 videos with 4021 code-mixed Hindi-English utterances collected from YouTube. Each utterance within this dataset has been meticulously annotated for toxicity, severity, and sentiment labels. We have developed an advanced Multimodal Multitask framework built for Toxicity detection in Video Content by leveraging Large Language Models (LLMs), crafted for the primary objective along with the additional tasks of conducting sentiment and severity analysis. ToxVidLLM incorporates three key modules the Encoder module, Cross-Modal Synchronization module, and Multitask module crafting a generic multimodal LLM customized for intricate video classification tasks. Our experiments reveal that incorporating multiple modalities from the videos substantially enhances the performance of toxic content detection by achieving an Accuracy and Weighted F1 score of 94.29% and 94.35%, respectively.
Authors: Yukai Zhang, Ao Xu, Zihao Li, Tieru Wu
Abstract: In the realm of Artificial Intelligence (AI), the importance of Explainable Artificial Intelligence (XAI) is increasingly recognized, particularly as AI models become more integral to our lives. One notable single-instance XAI approach is counterfactual explanation, which aids users in comprehending a model's decisions and offers guidance on altering these decisions. Specifically in the context of image classification models, effective image counterfactual explanations can significantly enhance user understanding. This paper introduces a novel method for computing feature importance within the feature space of a black-box model. By employing information fusion techniques, our method maximizes the use of data to address feature counterfactual explanations in the feature space. Subsequently, we utilize an image generation model to transform these feature counterfactual explanations into image counterfactual explanations. Our experiments demonstrate that the counterfactual explanations generated by our method closely resemble the original images in both pixel and feature spaces. Additionally, our method outperforms established baselines, achieving impressive experimental results.
Authors: Ruyi Zha, Tao Jun Lin, Yuanhao Cai, Jiwen Cao, Yanhao Zhang, Hongdong Li
Abstract: 3D Gaussian splatting (3DGS) has shown promising results in image rendering and surface reconstruction. However, its potential in volumetric reconstruction tasks, such as X-ray computed tomography, remains under-explored. This paper introduces R2-Gaussian, the first 3DGS-based framework for sparse-view tomographic reconstruction. By carefully deriving X-ray rasterization functions, we discover a previously unknown integration bias in the standard 3DGS formulation, which hampers accurate volume retrieval. To address this issue, we propose a novel rectification technique via refactoring the projection from 3D to 2D Gaussians. Our new method presents three key innovations: (1) introducing tailored Gaussian kernels, (2) extending rasterization to X-ray imaging, and (3) developing a CUDA-based differentiable voxelizer. Extensive experiments demonstrate that our method outperforms state-of-the-art approaches by 0.93 dB in PSNR and 0.014 in SSIM. Crucially, it delivers high-quality results in 3 minutes, which is 12x faster than NeRF-based methods and on par with traditional algorithms. The superior performance and rapid convergence of our method highlight its practical value.
Authors: Christina Winkler, Paula Harder, David Rolnick
Abstract: Predictions of global climate models typically operate on coarse spatial scales due to the large computational costs of climate simulations. This has led to a considerable interest in methods for statistical downscaling, a similar process to super-resolution in the computer vision context, to provide more local and regional climate information. In this work, we apply conditional normalizing flows to the task of climate variable downscaling. We showcase its successful performance on an ERA5 water content dataset for different upsampling factors. Additionally, we show that the method allows us to assess the predictive uncertainty in terms of standard deviation from the fitted conditional distribution mean.
Authors: Wenbo Yu, Hao Fang, Bin Chen, Xiaohang Sui, Chuan Chen, Hao Wu, Shu-Tao Xia, Ke Xu
Abstract: Gradient Inversion Attacks invert the transmitted gradients in Federated Learning (FL) systems to reconstruct the sensitive data of local clients and have raised considerable privacy concerns. A majority of gradient inversion methods rely heavily on explicit prior knowledge (e.g., a well pre-trained generative model), which is often unavailable in realistic scenarios. To alleviate this issue, researchers have proposed to leverage the implicit prior knowledge of an over-parameterized network. However, they only utilize a fixed neural architecture for all the attack settings. This would hinder the adaptive use of implicit architectural priors and consequently limit the generalizability. In this paper, we further exploit such implicit prior knowledge by proposing Gradient Inversion via Neural Architecture Search (GI-NAS), which adaptively searches the network and captures the implicit priors behind neural architectures. Extensive experiments verify that our proposed GI-NAS can achieve superior attack performance compared to state-of-the-art gradient inversion methods, even under more practical settings with high-resolution images, large-sized batches, and advanced defense strategies.
Authors: Chao Wang, Giulio Franzese, Alessandro Finamore, Massimo Gallo, Pietro Michiardi
Abstract: Diffusion models for Text-to-Image (T2I) conditional generation have seen tremendous success recently. Despite their success, accurately capturing user intentions with these models still requires a laborious trial and error process. This challenge is commonly identified as a model alignment problem, an issue that has attracted considerable attention by the research community. Instead of relying on fine-grained linguistic analyses of prompts, human annotation, or auxiliary vision-language models to steer image generation, in this work we present a novel method that relies on an information-theoretic alignment measure. In a nutshell, our method uses self-supervised fine-tuning and relies on point-wise mutual information between prompts and images to define a synthetic training set to induce model alignment. Our comparative analysis shows that our method is on-par or superior to the state-of-the-art, yet requires nothing but a pre-trained denoising network to estimate MI and a lightweight fine-tuning strategy.
Authors: Jingwei Li, Jing Dong, Tianxing He, Jingzhao Zhang
Abstract: Identifying whether an artwork was used to train a diffusion model is an important research topic, given the rising popularity of AI-generated art and the associated copyright concerns. The work approaches this problem from the membership inference attack (MIA) perspective. We first identify the limitations of applying existing MIA methods for copyright protection: the required access of internal U-nets and the choice of non-member datasets for evaluation. To address the above problems, we introduce a novel black-box membership inference attack method that operates without needing access to the model's internal U-net. We then construct a DALL-E generated dataset for a more comprehensive evaluation. We validate our method across various setups, and our experimental results outperform previous works.
Authors: Linus Ericsson, Miguel Espinosa, Chenhongyi Yang, Antreas Antoniou, Amos Storkey, Shay B. Cohen, Steven McDonagh, Elliot J. Crowley
Abstract: Neural architecture search (NAS) finds high performing networks for a given task. Yet the results of NAS are fairly prosaic; they did not e.g. create a shift from convolutional structures to transformers. This is not least because the search spaces in NAS often aren't diverse enough to include such transformations a priori. Instead, for NAS to provide greater potential for fundamental design shifts, we need a novel expressive search space design which is built from more fundamental operations. To this end, we introduce einspace, a search space based on a parameterised probabilistic context-free grammar. Our space is versatile, supporting architectures of various sizes and complexities, while also containing diverse network operations which allow it to model convolutions, attention components and more. It contains many existing competitive architectures, and provides flexibility for discovering new ones. Using this search space, we perform experiments to find novel architectures as well as improvements on existing ones on the diverse Unseen NAS datasets. We show that competitive architectures can be obtained by searching from scratch, and we consistently find large improvements when initialising the search with strong baselines. We believe that this work is an important advancement towards a transformative NAS paradigm where search space expressivity and strategic search initialisation play key roles.
Authors: Ming Du, Tao Zhou, Junjing Deng, Daniel J. Ching, Steven Henke, Mathew J. Cherukara
Abstract: Ptychography is a powerful imaging technique that is used in a variety of fields, including materials science, biology, and nanotechnology. However, the accuracy of the reconstructed ptychography image is highly dependent on the accuracy of the recorded probe positions which often contain errors. These errors are typically corrected jointly with phase retrieval through numerical optimization approaches. When the error accumulates along the scan path or when the error magnitude is large, these approaches may not converge with satisfactory result. We propose a fundamentally new approach for ptychography probe position prediction for data with large position errors, where a neural network is used to make single-shot phase retrieval on individual diffraction patterns, yielding the object image at each scan point. The pairwise offsets among these images are then found using a robust image registration method, and the results are combined to yield the complete scan path by constructing and solving a linear equation. We show that our method can achieve good position prediction accuracy for data with large and accumulating errors on the order of $10^2$ pixels, a magnitude that often makes optimization-based algorithms fail to converge. For ptychography instruments without sophisticated position control equipment such as interferometers, our method is of significant practical potential.
Authors: Metod Jazbec, Alexander Timans, Tin Had\v{z}i Veljkovi\'c, Kaspar Sakmann, Dan Zhang, Christian A. Naesseth, Eric Nalisnick
Abstract: Scaling machine learning models significantly improves their performance. However, such gains come at the cost of inference being slow and resource-intensive. Early-exit neural networks (EENNs) offer a promising solution: they accelerate inference by allowing intermediate layers to exit and produce a prediction early. Yet a fundamental issue with EENNs is how to determine when to exit without severely degrading performance. In other words, when is it 'safe' for an EENN to go 'fast'? To address this issue, we investigate how to adapt frameworks of risk control to EENNs. Risk control offers a distribution-free, post-hoc solution that tunes the EENN's exiting mechanism so that exits only occur when the output is of sufficient quality. We empirically validate our insights on a range of vision and language tasks, demonstrating that risk control can produce substantial computational savings, all the while preserving user-specified performance goals.
Authors: Siddarth Venkatraman, Moksh Jain, Luca Scimeca, Minsu Kim, Marcin Sendera, Mohsin Hasan, Luke Rowe, Sarthak Mittal, Pablo Lemos, Emmanuel Bengio, Alexandre Adam, Jarrid Rector-Brooks, Yoshua Bengio, Glen Berseth, Nikolay Malkin
Abstract: Diffusion models have emerged as effective distribution estimators in vision, language, and reinforcement learning, but their use as priors in downstream tasks poses an intractable posterior inference problem. This paper studies amortized sampling of the posterior over data, $\mathbf{x}\sim p^{\rm post}(\mathbf{x})\propto p(\mathbf{x})r(\mathbf{x})$, in a model that consists of a diffusion generative model prior $p(\mathbf{x})$ and a black-box constraint or likelihood function $r(\mathbf{x})$. We state and prove the asymptotic correctness of a data-free learning objective, relative trajectory balance, for training a diffusion model that samples from this posterior, a problem that existing methods solve only approximately or in restricted cases. Relative trajectory balance arises from the generative flow network perspective on diffusion models, which allows the use of deep reinforcement learning techniques to improve mode coverage. Experiments illustrate the broad potential of unbiased inference of arbitrary posteriors under diffusion priors: in vision (classifier guidance), language (infilling under a discrete diffusion LLM), and multimodal data (text-to-image generation). Beyond generative modeling, we apply relative trajectory balance to the problem of continuous control with a score-based behavior prior, achieving state-of-the-art results on benchmarks in offline reinforcement learning.
Authors: Matej Gazda, Samuel Kadoury, Jakub Gazda, Peter Drotar
Abstract: Transthoracic Echocardiography (TTE) is a fundamental, non-invasive diagnostic tool in cardiovascular medicine, enabling detailed visualization of cardiac structures crucial for diagnosing various heart conditions. Despite its widespread use, TTE ultrasound imaging faces inherent limitations, notably the trade-off between field of view (FoV) and resolution. This paper introduces a novel application of conditional Generative Adversarial Networks (cGANs), specifically designed to extend the FoV in TTE ultrasound imaging while maintaining high resolution. Our proposed cGAN architecture, termed echoGAN, demonstrates the capability to generate realistic anatomical structures through outpainting, effectively broadening the viewable area in medical imaging. This advancement has the potential to enhance both automatic and manual ultrasound navigation, offering a more comprehensive view that could significantly reduce the learning curve associated with ultrasound imaging and aid in more accurate diagnoses. The results confirm that echoGAN reliably reproduce detailed cardiac features, thereby promising a significant step forward in the field of non-invasive cardiac naviagation and diagnostics.
Authors: Linlin Yu, Bowen Yang, Tianhao Wang, Kangshuo Li, Feng Chen
Abstract: The fusion of raw features from multiple sensors on an autonomous vehicle to create a Bird's Eye View (BEV) representation is crucial for planning and control systems. There is growing interest in using deep learning models for BEV semantic segmentation. Anticipating segmentation errors and improving the explainability of DNNs is essential for autonomous driving, yet it is under-studied. This paper introduces a benchmark for predictive uncertainty quantification in BEV segmentation. The benchmark assesses various approaches across three popular datasets using two representative backbones and focuses on the effectiveness of predicted uncertainty in identifying misclassified and out-of-distribution (OOD) pixels, as well as calibration. Empirical findings highlight the challenges in uncertainty quantification. Our results find that evidential deep learning based approaches show the most promise by efficiently quantifying aleatoric and epistemic uncertainty. We propose the Uncertainty-Focal-Cross-Entropy (UFCE) loss, designed for highly imbalanced data, which consistently improves the segmentation quality and calibration. Additionally, we introduce a vacuity-scaled regularization term that enhances the model's focus on high uncertainty pixels, improving epistemic uncertainty quantification.
Authors: Zhen Qin, Yuxin Mao, Xuyang Shen, Dong Li, Jing Zhang, Yuchao Dai, Yiran Zhong
Abstract: Linear attention mechanisms have gained prominence in causal language models due to their linear computational complexity and enhanced speed. However, the inherent decay mechanism in linear attention presents challenges when applied to multi-dimensional sequence modeling tasks, such as image processing and multi-modal learning. In these scenarios, the utilization of sequential scanning to establish a global receptive field necessitates multiple scans for multi-dimensional data, thereby leading to inefficiencies. This paper identifies the inefficiency caused by a multiplicative linear recurrence and proposes an efficient alternative additive linear recurrence to avoid the issue, as it can handle multi-dimensional data within a single scan. We further develop an efficient multi-dimensional sequential modeling framework called LightNet based on the new recurrence. Moreover, we present two new multi-dimensional linear relative positional encoding methods, MD-TPE and MD-LRPE to enhance the model's ability to discern positional information in multi-dimensional scenarios. Our empirical evaluations across various tasks, including image classification, image generation, bidirectional language modeling, and autoregressive language modeling, demonstrate the efficacy of LightNet, showcasing its potential as a versatile and efficient solution for multi-dimensional sequential modeling.
Authors: Deng Cao, Hongbo Zhang, Rajveer Dhillon
Abstract: Organic weed control is a vital to improve crop yield with a sustainable approach. In this work, a directed energy weed control robot prototype specifically designed for organic farms is proposed. The robot uses a novel distributed array robot (DAR) unit for weed treatment. Soybean and corn databases are built to train deep learning neural nets to perform weed recognition. The initial deep learning neural nets show a high performance in classifying crops. The robot uses a patented directed energy plant eradication recipe that is completely organic and UV-C free, with no chemical damage or physical disturbance to the soil. The deep learning can classify 8 common weed species in a soybean field under natural environment with up to 98% accuracy.
Authors: Jiangwei Zhao, Zejia Liu, Xiaohan Guo, Lili Pan
Abstract: Disentanglement, a critical concern in interpretable machine learning, has also garnered significant attention from the computer vision community. Many existing GAN-based class disentanglement (unsupervised) approaches, such as InfoGAN and its variants, primarily aim to maximize the mutual information (MI) between the generated image and its latent codes. However, this focus may lead to a tendency for the network to generate highly similar images when presented with the same latent class factor, potentially resulting in mode collapse or mode dropping. To alleviate this problem, we propose \texttt{CoDeGAN} (Contrastive Disentanglement for Generative Adversarial Networks), where we relax similarity constraints for disentanglement from the image domain to the feature domain. This modification not only enhances the stability of GAN training but also improves their disentangling capabilities. Moreover, we integrate self-supervised pre-training into CoDeGAN to learn semantic representations, significantly facilitating unsupervised disentanglement. Extensive experimental results demonstrate the superiority of our method over state-of-the-art approaches across multiple benchmarks. The code is available at https://github.com/learninginvision/CoDeGAN.
Authors: Mengnan Shi, Chang Liu, Jianbin Jiao, Qixiang Ye
Abstract: Gating modules have been widely explored in dynamic network pruning to reduce the run-time computational cost of deep neural networks while preserving the representation of features. Despite the substantial progress, existing methods remain ignoring the consistency between feature and gate distributions, which may lead to distortion of gated features. In this paper, we propose a feature-gate coupling (FGC) approach aiming to align distributions of features and gates. FGC is a plug-and-play module, which consists of two steps carried out in an iterative self-supervised manner. In the first step, FGC utilizes the $k$-Nearest Neighbor method in the feature space to explore instance neighborhood relationships, which are treated as self-supervisory signals. In the second step, FGC exploits contrastive learning to regularize gating modules with generated self-supervisory signals, leading to the alignment of instance neighborhood relationships within the feature and gate spaces. Experimental results validate that the proposed FGC method improves the baseline approach with significant margins, outperforming the state-of-the-arts with better accuracy-computation trade-off. Code is publicly available.
Authors: Jiaming Zhang, Kailun Yang, Hao Shi, Simon Rei{\ss}, Kunyu Peng, Chaoxiang Ma, Haodong Fu, Philip H. S. Torr, Kaiwei Wang, Rainer Stiefelhagen
Abstract: In this paper, we address panoramic semantic segmentation which is under-explored due to two critical challenges: (1) image distortions and object deformations on panoramas; (2) lack of semantic annotations in the 360{\deg} imagery. To tackle these problems, first, we propose the upgraded Transformer for Panoramic Semantic Segmentation, i.e., Trans4PASS+, equipped with Deformable Patch Embedding (DPE) and Deformable MLP (DMLPv2) modules for handling object deformations and image distortions whenever (before or after adaptation) and wherever (shallow or deep levels). Second, we enhance the Mutual Prototypical Adaptation (MPA) strategy via pseudo-label rectification for unsupervised domain adaptive panoramic segmentation. Third, aside from Pinhole-to-Panoramic (Pin2Pan) adaptation, we create a new dataset (SynPASS) with 9,080 panoramic images, facilitating Synthetic-to-Real (Syn2Real) adaptation scheme in 360{\deg} imagery. Extensive experiments are conducted, which cover indoor and outdoor scenarios, and each of them is investigated with Pin2Pan and Syn2Real regimens. Trans4PASS+ achieves state-of-the-art performances on four domain adaptive panoramic semantic segmentation benchmarks. Code is available at https://github.com/jamycheung/Trans4PASS.
Authors: Sunil Hwang, Jaehong Yoon, Youngwan Lee, Sung Ju Hwang
Abstract: Masked Video Autoencoder (MVA) approaches have demonstrated their potential by significantly outperforming previous video representation learning methods. However, they waste an excessive amount of computations and memory in predicting uninformative tokens/frames due to random masking strategies. (e.g., over 16 nodes with 128 NVIDIA A100 GPUs). To resolve this issue, we exploit the unequal information density among the patches in videos and propose EVEREST, a surprisingly efficient MVA approach for video representation learning that finds tokens containing rich motion features and discards uninformative ones during both pre-training and fine-tuning. We further present an information-intensive frame selection strategy that allows the model to focus on informative and causal frames with minimal redundancy. Our method significantly reduces the computation and memory requirements of MVA, enabling the pre-training and fine-tuning on a single machine with 8 GPUs while achieving comparable performance to computation- and memory-heavy baselines on multiple benchmarks and the uncurated Ego4D dataset. We hope that our work contributes to reducing the barrier to further research on video understanding.
Authors: Hubert Leterme, K\'evin Polisano, Val\'erie Perrier, Karteek Alahari
Abstract: We propose a novel method to increase shift invariance and prediction accuracy in convolutional neural networks. Specifically, we replace the first-layer combination "real-valued convolutions + max pooling" (RMax) by "complex-valued convolutions + modulus" (CMod), which is stable to translations, or shifts. To justify our approach, we claim that CMod and RMax produce comparable outputs when the convolution kernel is band-pass and oriented (Gabor-like filter). In this context, CMod can therefore be considered as a stable alternative to RMax. To enforce this property, we constrain the convolution kernels to adopt such a Gabor-like structure. The corresponding architecture is called mathematical twin, because it employs a well-defined mathematical operator to mimic the behavior of the original, freely-trained model. Our approach achieves superior accuracy on ImageNet and CIFAR-10 classification tasks, compared to prior methods based on low-pass filtering. Arguably, our approach's emphasis on retaining high-frequency details contributes to a better balance between shift invariance and information preservation, resulting in improved performance. Furthermore, it has a lower computational cost and memory footprint than concurrent work, making it a promising solution for practical implementation.
Authors: Meng You, Junhui Hou
Abstract: The challenge of dynamic view synthesis from dynamic monocular videos, i.e., synthesizing novel views for free viewpoints given a monocular video of a dynamic scene captured by a moving camera, mainly lies in accurately modeling the \textbf{dynamic objects} of a scene using limited 2D frames, each with a varying timestamp and viewpoint. Existing methods usually require pre-processed 2D optical flow and depth maps by off-the-shelf methods to supervise the network, making them suffer from the inaccuracy of the pre-processed supervision and the ambiguity when lifting the 2D information to 3D. In this paper, we tackle this challenge in an unsupervised fashion. Specifically, we decouple the motion of the dynamic objects into object motion and camera motion, respectively regularized by proposed unsupervised surface consistency and patch-based multi-view constraints. The former enforces the 3D geometric surfaces of moving objects to be consistent over time, while the latter regularizes their appearances to be consistent across different viewpoints. Such a fine-grained motion formulation can alleviate the learning difficulty for the network, thus enabling it to produce not only novel views with higher quality but also more accurate scene flows and depth than existing methods requiring extra supervision.
Authors: Tianshu Yu, Changqun Xia, Jia Li
Abstract: Video portrait segmentation (VPS), aiming at segmenting prominent foreground portraits from video frames, has received much attention in recent years. However, simplicity of existing VPS datasets leads to a limitation on extensive research of the task. In this work, we propose a new intricate large-scale Multi-scene Video Portrait Segmentation dataset MVPS consisting of 101 video clips in 7 scenario categories, in which 10,843 sampled frames are finely annotated at pixel level. The dataset has diverse scenes and complicated background environments, which is the most complex dataset in VPS to our best knowledge. Through the observation of a large number of videos with portraits during dataset construction, we find that due to the joint structure of human body, motion of portraits is part-associated, which leads that different parts are relatively independent in motion. That is, motion of different parts of the portraits is imbalanced. Towards this imbalance, an intuitive and reasonable idea is that different motion states in portraits can be better exploited by decoupling the portraits into parts. To achieve this, we propose a Part-Decoupling Network (PDNet) for video portrait segmentation. Specifically, an Inter-frame Part-Discriminated Attention (IPDA) module is proposed which unsupervisedly segments portrait into parts and utilizes different attentiveness on discriminative features specified to each different part. In this way, appropriate attention can be imposed to portrait parts with imbalanced motion to extract part-discriminated correlations, so that the portraits can be segmented more accurately. Experimental results demonstrate that our method achieves leading performance with the comparison to state-of-the-art methods.
Authors: Jung H. Lee, Sujith Vijayan
Abstract: Deep learning models loosely mimic bottom-up signal pathways from low-order sensory areas to high-order cognitive areas. After training, DL models can outperform humans on some domain-specific tasks, but their decision-making process has been known to be easily disrupted. Since the human brain consists of multiple functional areas highly connected to one another and relies on intricate interplays between bottom-up and top-down (from high-order to low-order areas) processing, we hypothesize that incorporating top-down signal processing may make DL models more robust. To address this hypothesis, we propose a certification process mimicking selective attention and test if it could make DL models more robust. Our empirical evaluations suggest that this newly proposed certification can improve DL models' accuracy and help us build safety measures to alleviate their vulnerabilities with both artificial and natural adversarial examples.
Authors: Ayush Sarkar, Hanlin Mai, Amitabh Mahapatra, Svetlana Lazebnik, D. A. Forsyth, Anand Bhattad
Abstract: Generative models can produce impressively realistic images. This paper demonstrates that generated images have geometric features different from those of real images. We build a set of collections of generated images, prequalified to fool simple, signal-based classifiers into believing they are real. We then show that prequalified generated images can be identified reliably by classifiers that only look at geometric properties. We use three such classifiers. All three classifiers are denied access to image pixels, and look only at derived geometric features. The first classifier looks at the perspective field of the image, the second looks at lines detected in the image, and the third looks at relations between detected objects and shadows. Our procedure detects generated images more reliably than SOTA local signal based detectors, for images from a number of distinct generators. Saliency maps suggest that the classifiers can identify geometric problems reliably. We conclude that current generators cannot reliably reproduce geometric properties of real images.
Authors: Huajian Huang, Changkun Liu, Yipeng Zhu, Hui Cheng, Tristan Braud, Sai-Kit Yeung
Abstract: Portable 360$^\circ$ cameras are becoming a cheap and efficient tool to establish large visual databases. By capturing omnidirectional views of a scene, these cameras could expedite building environment models that are essential for visual localization. However, such an advantage is often overlooked due to the lack of valuable datasets. This paper introduces a new benchmark dataset, 360Loc, composed of 360$^\circ$ images with ground truth poses for visual localization. We present a practical implementation of 360$^\circ$ mapping combining 360$^\circ$ images with lidar data to generate the ground truth 6DoF poses. 360Loc is the first dataset and benchmark that explores the challenge of cross-device visual positioning, involving 360$^\circ$ reference frames, and query frames from pinhole, ultra-wide FoV fisheye, and 360$^\circ$ cameras. We propose a virtual camera approach to generate lower-FoV query frames from 360$^\circ$ images, which ensures a fair comparison of performance among different query types in visual localization tasks. We also extend this virtual camera approach to feature matching-based and pose regression-based methods to alleviate the performance loss caused by the cross-device domain gap, and evaluate its effectiveness against state-of-the-art baselines. We demonstrate that omnidirectional visual localization is more robust in challenging large-scale scenes with symmetries and repetitive structures. These results provide new insights into 360-camera mapping and omnidirectional visual localization with cross-device queries.
Authors: R\'emi Kazmierczak, Elo\"ise Berthier, Goran Frehse, Gianni Franchi
Abstract: In this paper, we introduce an explainable algorithm designed from a multi-modal foundation model, that performs fast and explainable image classification. Drawing inspiration from CLIP-based Concept Bottleneck Models (CBMs), our method creates a latent space where each neuron is linked to a specific word. Observing that this latent space can be modeled with simple distributions, we use a Mixture of Gaussians (MoG) formalism to enhance the interpretability of this latent space. Then, we introduce CLIP-QDA, a classifier that only uses statistical values to infer labels from the concepts. In addition, this formalism allows for both local and global explanations. These explanations come from the inner design of our architecture, our work is part of a new family of greybox models, combining performances of opaque foundation models and the interpretability of transparent models. Our empirical findings show that in instances where the MoG assumption holds, CLIP-QDA achieves similar accuracy with state-of-the-art methods CBMs. Our explanations compete with existing XAI methods while being faster to compute.
Authors: Gu\'enol\'e Fiche, Simon Leglaive, Xavier Alameda-Pineda, Antonio Agudo, Francesc Moreno-Noguer
Abstract: Previous works on Human Pose and Shape Estimation (HPSE) from RGB images can be broadly categorized into two main groups: parametric and non-parametric approaches. Parametric techniques leverage a low-dimensional statistical body model for realistic results, whereas recent non-parametric methods achieve higher precision by directly regressing the 3D coordinates of the human body mesh. This work introduces a novel paradigm to address the HPSE problem, involving a low-dimensional discrete latent representation of the human mesh and framing HPSE as a classification task. Instead of predicting body model parameters or 3D vertex coordinates, we focus on predicting the proposed discrete latent representation, which can be decoded into a registered human mesh. This innovative paradigm offers two key advantages. Firstly, predicting a low-dimensional discrete representation confines our predictions to the space of anthropomorphic poses and shapes even when little training data is available. Secondly, by framing the problem as a classification task, we can harness the discriminative power inherent in neural networks. The proposed model, VQ-HPS, predicts the discrete latent representation of the mesh. The experimental results demonstrate that VQ-HPS outperforms the current state-of-the-art non-parametric approaches while yielding results as realistic as those produced by parametric methods when trained with little data. VQ-HPS also shows promising results when training on large-scale datasets, highlighting the significant potential of the classification approach for HPSE. See the project page at https://g-fiche.github.io/research-pages/vqhps/
Authors: Zhilu Zhang, Shuohao Zhang, Renlong Wu, Zifei Yan, Wangmeng Zuo
Abstract: It is highly desired but challenging to acquire high-quality photos with clear content in low-light environments. Although multi-image processing methods (using burst, dual-exposure, or multi-exposure images) have made significant progress in addressing this issue, they typically focus on specific restoration or enhancement problems, and do not fully explore the potential of utilizing multiple images. Motivated by the fact that multi-exposure images are complementary in denoising, deblurring, high dynamic range imaging, and super-resolution, we propose to utilize exposure bracketing photography to unify image restoration and enhancement tasks in this work. Due to the difficulty in collecting real-world pairs, we suggest a solution that first pre-trains the model with synthetic paired data and then adapts it to real-world unlabeled images. In particular, a temporally modulated recurrent network (TMRNet) and self-supervised adaptation method are proposed. Moreover, we construct a data simulation pipeline to synthesize pairs and collect real-world images from 200 nighttime scenarios. Experiments on both datasets show that our method performs favorably against the state-of-the-art multi-image processing ones. The dataset, code, and pre-trained models are available at https://github.com/cszhilu1998/BracketIRE.
Authors: Chintan Trivedi, Nemanja Ra\v{s}ajski, Konstantinos Makantasis, Antonios Liapis, Georgios N. Yannakakis
Abstract: Domain randomization is an effective computer vision technique for improving transferability of vision models across visually distinct domains exhibiting similar content. Existing approaches, however, rely extensively on tweaking complex and specialized simulation engines that are difficult to construct, subsequently affecting their feasibility and scalability. This paper introduces BehAVE, a video understanding framework that uniquely leverages the plethora of existing commercial video games for domain randomization, without requiring access to their simulation engines. Under BehAVE (1) the inherent rich visual diversity of video games acts as the source of randomization and (2) player behavior -- represented semantically via textual descriptions of actions -- guides the *alignment* of videos with similar content. We test BehAVE on 25 games of the first-person shooter (FPS) genre across various video and text foundation models and we report its robustness for domain randomization. BehAVE successfully aligns player behavioral patterns and is able to zero-shot transfer them to multiple unseen FPS games when trained on just one FPS game. In a more challenging setting, BehAVE manages to improve the zero-shot transferability of foundation models to unseen FPS games (up to 22%) even when trained on a game of a different genre (Minecraft). Code and dataset can be found at https://github.com/nrasajski/BehAVE.
Authors: Ivana Bala\v{z}evi\'c, Yuge Shi, Pinelopi Papalampidi, Rahma Chaabouni, Skanda Koppula, Olivier J. H\'enaff
Abstract: Most transformer-based video encoders are limited to short temporal contexts due to their quadratic complexity. While various attempts have been made to extend this context, this has often come at the cost of both conceptual and computational complexity. We propose to instead re-purpose existing pre-trained video transformers by simply fine-tuning them to attend to memories derived non-parametrically from past activations. By leveraging redundancy reduction, our memory-consolidated vision transformer (MC-ViT) effortlessly extends its context far into the past and exhibits excellent scaling behavior when learning from longer videos. In doing so, MC-ViT sets a new state-of-the-art in long-context video understanding on EgoSchema, Perception Test, and Diving48, outperforming methods that benefit from orders of magnitude more parameters.
Authors: Parisa Farmanifard, Arun Ross
Abstract: Iris segmentation is a critical component of an iris biometric system and it involves extracting the annular iris region from an ocular image. In this work, we develop a pixel-level iris segmentation model from a foundational model, viz., Segment Anything Model (SAM), that has been successfully used for segmenting arbitrary objects. The primary contribution of this work lies in the integration of different loss functions during the fine-tuning of SAM on ocular images. In particular, the importance of Focal Loss is borne out in the fine-tuning process since it strategically addresses the class imbalance problem (i.e., iris versus non-iris pixels). Experiments on ND-IRIS-0405, CASIA-Iris-Interval-v3, and IIT-Delhi-Iris datasets convey the efficacy of the trained model for the task of iris segmentation. For instance, on the ND-IRIS-0405 dataset, an average segmentation accuracy of 99.58% was achieved, compared to the best baseline performance of 89.75%.
Authors: Jihyung Kil, Farideh Tavazoee, Dongyeop Kang, Joo-Kyung Kim
Abstract: Visual Question Answering (VQA) often involves diverse reasoning scenarios across Vision and Language (V&L). Most prior VQA studies, however, have merely focused on assessing the model's overall accuracy without evaluating it on different reasoning cases. Furthermore, some recent works observe that conventional Chain-of-Thought (CoT) prompting fails to generate effective reasoning for VQA, especially for complex scenarios requiring multi-hop reasoning. In this paper, we propose II-MMR, a novel idea to identify and improve multi-modal multi-hop reasoning in VQA. In specific, II-MMR takes a VQA question with an image and finds a reasoning path to reach its answer using two novel language promptings: (i) answer prediction-guided CoT prompt, or (ii) knowledge triplet-guided prompt. II-MMR then analyzes this path to identify different reasoning cases in current VQA benchmarks by estimating how many hops and what types (i.e., visual or beyond-visual) of reasoning are required to answer the question. On popular benchmarks including GQA and A-OKVQA, II-MMR observes that most of their VQA questions are easy to answer, simply demanding "single-hop" reasoning, whereas only a few questions require "multi-hop" reasoning. Moreover, while the recent V&L model struggles with such complex multi-hop reasoning questions even using the traditional CoT method, II-MMR shows its effectiveness across all reasoning cases in both zero-shot and fine-tuning settings.
Authors: James Oldfield, Markos Georgopoulos, Grigorios G. Chrysos, Christos Tzelepis, Yannis Panagakis, Mihalis A. Nicolaou, Jiankang Deng, Ioannis Patras
Abstract: The Mixture of Experts (MoE) paradigm provides a powerful way to decompose dense layers into smaller, modular computations often more amenable to human interpretation, debugging, and editability. However, a major challenge lies in the computational cost of scaling the number of experts high enough to achieve fine-grained specialization. In this paper, we propose the Multilinear Mixture of Experts ($\mu$MoE) layer to address this, focusing on vision models. $\mu$MoE layers enable scalable expert specialization by performing an implicit computation on prohibitively large weight tensors entirely in factorized form. Consequently, $\mu$MoEs (1) avoid the restrictively high inference-time costs of 'soft' MoEs, yet (2) do not inherit the training issues of the popular 'sparse' MoEs' discrete (non-differentiable) expert routing. We present both qualitative and quantitative evidence that scaling $\mu$MoE layers when fine-tuning foundation models for vision tasks leads to more specialized experts at the class-level, further enabling manual bias correction in CelebA attribute classification. Finally, we show qualitative results demonstrating the expert specialism achieved when pre-training large GPT2 and MLP-Mixer models with parameter-matched $\mu$MoE blocks at every layer, maintaining comparable accuracy. Our code is available at: https://github.com/james-oldfield/muMoE.
Authors: Li Lin, Yixiang Liu, Jiewei Wu, Pujin Cheng, Zhiyuan Cai, Kenneth K. Y. Wong, Xiaoying Tang
Abstract: Federated learning (FL) effectively mitigates the data silo challenge brought about by policies and privacy concerns, implicitly harnessing more data for deep model training. However, traditional centralized FL models grapple with diverse multi-center data, especially in the face of significant data heterogeneity, notably in medical contexts. In the realm of medical image segmentation, the growing imperative to curtail annotation costs has amplified the importance of weakly-supervised techniques which utilize sparse annotations such as points, scribbles, etc. A pragmatic FL paradigm shall accommodate diverse annotation formats across different sites, which research topic remains under-investigated. In such context, we propose a novel personalized FL framework with learnable prompt and aggregation (FedLPPA) to uniformly leverage heterogeneous weak supervision for medical image segmentation. In FedLPPA, a learnable universal knowledge prompt is maintained, complemented by multiple learnable personalized data distribution prompts and prompts representing the supervision sparsity. Integrated with sample features through a dual-attention mechanism, those prompts empower each local task decoder to adeptly adjust to both the local distribution and the supervision form. Concurrently, a dual-decoder strategy, predicated on prompt similarity, is introduced for enhancing the generation of pseudo-labels in weakly-supervised learning, alleviating overfitting and noise accumulation inherent to local data, while an adaptable aggregation method is employed to customize the task decoder on a parameter-wise basis. Extensive experiments on four distinct medical image segmentation tasks involving different modalities underscore the superiority of FedLPPA, with its efficacy closely parallels that of fully supervised centralized training. Our code and data will be available.
Authors: Penghao Zhao, Hailin Zhang, Qinhan Yu, Zhengren Wang, Yunteng Geng, Fangcheng Fu, Ling Yang, Wentao Zhang, Jie Jiang, Bin Cui
Abstract: Advancements in model algorithms, the growth of foundational models, and access to high-quality datasets have propelled the evolution of Artificial Intelligence Generated Content (AIGC). Despite its notable successes, AIGC still faces hurdles such as updating knowledge, handling long-tail data, mitigating data leakage, and managing high training and inference costs. Retrieval-Augmented Generation (RAG) has recently emerged as a paradigm to address such challenges. In particular, RAG introduces the information retrieval process, which enhances the generation process by retrieving relevant objects from available data stores, leading to higher accuracy and better robustness. In this paper, we comprehensively review existing efforts that integrate RAG technique into AIGC scenarios. We first classify RAG foundations according to how the retriever augments the generator, distilling the fundamental abstractions of the augmentation methodologies for various retrievers and generators. This unified perspective encompasses all RAG scenarios, illuminating advancements and pivotal technologies that help with potential future progress. We also summarize additional enhancements methods for RAG, facilitating effective engineering and implementation of RAG systems. Then from another view, we survey on practical applications of RAG across different modalities and tasks, offering valuable references for researchers and practitioners. Furthermore, we introduce the benchmarks for RAG, discuss the limitations of current RAG systems, and suggest potential directions for future research. Github: https://github.com/PKU-DAIR/RAG-Survey.
Authors: Tung-Yu Wu, Sheng-Yu Huang, Yu-Chiang Frank Wang
Abstract: 3D visual grounding aims to identify the target object within a 3D point cloud scene referred to by a natural language description. Previous works usually require significant data relating to point color and their descriptions to exploit the corresponding complicated verbo-visual relations. In our work, we introduce Vigor, a novel Data-Efficient 3D Visual Grounding framework via Order-aware Referring. Vigor leverages LLM to produce a desirable referential order from the input description for 3D visual grounding. With the proposed stacked object-referring blocks, the predicted anchor objects in the above order allow one to locate the target object progressively without supervision on the identities of anchor objects or exact relations between anchor/target objects. In addition, we present an order-aware warm-up training strategy, which augments referential orders for pre-training the visual grounding framework. This allows us to better capture the complex verbo-visual relations and benefit the desirable data-efficient learning scheme. Experimental results on the NR3D and ScanRefer datasets demonstrate our superiority in low-resource scenarios. In particular, Vigor surpasses current state-of-the-art frameworks by 9.3% and 7.6% grounding accuracy under 1% data and 10% data settings on the NR3D dataset, respectively.
Authors: Yiwen Tang, Ray Zhang, Jiaming Liu, Zoey Guo, Dong Wang, Zhigang Wang, Bin Zhao, Shanghang Zhang, Peng Gao, Hongsheng Li, Xuelong Li
Abstract: Large foundation models have recently emerged as a prominent focus of interest, attaining superior performance in widespread scenarios. Due to the scarcity of 3D data, many efforts have been made to adapt pre-trained transformers from vision to 3D domains. However, such 2D-to-3D approaches are still limited, due to the potential loss of spatial geometries and high computation cost. More importantly, their frameworks are mainly designed for 2D models, lacking a general any-to-3D paradigm. In this paper, we introduce Any2Point, a parameter-efficient method to empower any-modality large models (vision, language, audio) for 3D understanding. Given a frozen transformer from any source modality, we propose a 3D-to-any (1D or 2D) virtual projection strategy that correlates the input 3D points to the original 1D or 2D positions within the source modality. This mechanism enables us to assign each 3D token with a positional encoding paired with the pre-trained model, which avoids 3D geometry loss caused by the true projection and better motivates the transformer for 3D learning with 1D/2D positional priors. Then, within each transformer block, we insert an any-to-3D guided adapter module for parameter-efficient fine-tuning. The adapter incorporates prior spatial knowledge from the source modality to guide the local feature aggregation of 3D tokens, compelling the semantic adaption of any-modality transformers. We conduct extensive experiments to showcase the effectiveness and efficiency of our method. Code and models are released at https://github.com/Ivan-Tang-3D/Any2Point.
Authors: Zixuan Zhu, Rui Wang, Cong Zou, Lihua Jing
Abstract: Recently, backdoor attacks have posed a serious security threat to the training process of deep neural networks (DNNs). The attacked model behaves normally on benign samples but outputs a specific result when the trigger is present. However, compared with the rocketing progress of backdoor attacks, existing defenses are difficult to deal with these threats effectively or require benign samples to work, which may be unavailable in real scenarios. In this paper, we find that the poisoned samples and benign samples can be distinguished with prediction entropy. This inspires us to propose a novel dual-network training framework: The Victim and The Beneficiary (V&B), which exploits a poisoned model to train a clean model without extra benign samples. Firstly, we sacrifice the Victim network to be a powerful poisoned sample detector by training on suspicious samples. Secondly, we train the Beneficiary network on the credible samples selected by the Victim to inhibit backdoor injection. Thirdly, a semi-supervised suppression strategy is adopted for erasing potential backdoors and improving model performance. Furthermore, to better inhibit missed poisoned samples, we propose a strong data augmentation method, AttentionMix, which works well with our proposed V&B framework. Extensive experiments on two widely used datasets against 6 state-of-the-art attacks demonstrate that our framework is effective in preventing backdoor injection and robust to various attacks while maintaining the performance on benign samples. Our code is available at https://github.com/Zixuan-Zhu/VaB.
Authors: Wenbin Guan, Zijiu Yang, Xiaohong Wu, Liqiong Chen, Feng Huang, Xiaohai He, Honggang Chen
Abstract: Presently, the task of few-shot object detection (FSOD) in remote sensing images (RSIs) has become a focal point of attention. Numerous few-shot detectors, particularly those based on two-stage detectors, face challenges when dealing with the multiscale complexities inherent in RSIs. Moreover, these detectors present impractical characteristics in real-world applications, mainly due to their unwieldy model parameters when handling large amount of data. In contrast, we recognize the advantages of one-stage detectors, including high detection speed and a global receptive field. Consequently, we choose the YOLOv7 one-stage detector as a baseline and subject it to a novel meta-learning training framework. This transformation allows the detector to adeptly address FSOD tasks while capitalizing on its inherent advantage of lightweight. Additionally, we thoroughly investigate the samples generated by the meta-learning strategy and introduce a novel meta-sampling approach to retain samples produced by our designed meta-detection head. Coupled with our devised meta-cross loss, we deliberately utilize "negative samples" that are often overlooked to extract valuable knowledge from them. This approach serves to enhance detection accuracy and efficiently refine the overall meta-learning strategy. To validate the effectiveness of our proposed detector, we conducted performance comparisons with current state-of-the-art detectors using the DIOR and NWPU VHR-10.v2 datasets, yielding satisfactory results.
Authors: Jian Liu, Wei Sun, Hui Yang, Zhiwen Zeng, Chongpei Liu, Jin Zheng, Xingyu Liu, Hossein Rahmani, Nicu Sebe, Ajmal Mian
Abstract: Object pose estimation is a fundamental computer vision problem with broad applications in augmented reality and robotics. Over the past decade, deep learning models, due to their superior accuracy and robustness, have increasingly supplanted conventional algorithms reliant on engineered point pair features. Nevertheless, several challenges persist in contemporary methods, including their dependency on labeled training data, model compactness, robustness under challenging conditions, and their ability to generalize to novel unseen objects. A recent survey discussing the progress made on different aspects of this area, outstanding challenges, and promising future directions, is missing. To fill this gap, we discuss the recent advances in deep learning-based object pose estimation, covering all three formulations of the problem, \emph{i.e.}, instance-level, category-level, and unseen object pose estimation. Our survey also covers multiple input data modalities, degrees-of-freedom of output poses, object properties, and downstream tasks, providing the readers with a holistic understanding of this field. Additionally, it discusses training paradigms of different domains, inference modes, application areas, evaluation metrics, and benchmark datasets, as well as reports the performance of current state-of-the-art methods on these benchmarks, thereby facilitating the readers in selecting the most suitable method for their application. Finally, the survey identifies key challenges, reviews the prevailing trends along with their pros and cons, and identifies promising directions for future research. We also keep tracing the latest works at https://github.com/CNJianLiu/Awesome-Object-Pose-Estimation.
URLs: https://github.com/CNJianLiu/Awesome-Object-Pose-Estimation.
Authors: Xinli Guo, Weidong Zhang, Ruonan Liu, Peng Han, Hongtian Chen
Abstract: With their high-fidelity scene representation capability, the attention of SLAM field is deeply attracted by the Neural Radiation Field (NeRF) and 3D Gaussian Splatting (3DGS). Recently, there has been a surge in NeRF-based SLAM, while 3DGS-based SLAM is sparse. A novel 3DGS-based SLAM approach with a fusion of deep visual feature, dual keyframe selection and 3DGS is presented in this paper. Compared with the existing methods, the proposed tracking is achieved by feature extraction and motion filter on each frame. The joint optimization of poses and 3D Gaussians runs through the entire mapping process. Additionally, the coarse-to-fine pose estimation and compact Gaussian scene representation are implemented by dual keyframe selection and novel loss functions. Experimental results demonstrate that the proposed algorithm not only outperforms the existing methods in tracking and mapping, but also has less memory usage.
Authors: Ying Jin, Pengyang Ling, Xiaoyi Dong, Pan Zhang, Jiaqi Wang, Dahua Lin
Abstract: Instruction-based image editing focuses on equipping a generative model with the capacity to adhere to human-written instructions for editing images. Current approaches typically comprehend explicit and specific instructions. However, they often exhibit a deficiency in executing active reasoning capacities required to comprehend instructions that are implicit or insufficiently defined. To enhance active reasoning capabilities and impart intelligence to the editing model, we introduce ReasonPix2Pix, a comprehensive reasoning-attentive instruction editing dataset. The dataset is characterized by 1) reasoning instruction, 2) more realistic images from fine-grained categories, and 3) increased variances between input and edited images. When fine-tuned with our dataset under supervised conditions, the model demonstrates superior performance in instructional editing tasks, independent of whether the tasks require reasoning or not. The code will be available at https://github.com/Jin-Ying/ReasonPix2Pix.
Authors: Dan Zhang, Jingjing Wang, Feng Luo
Abstract: In this paper, we present the Directly Denoising Diffusion Model (DDDM): a simple and generic approach for generating realistic images with few-step sampling, while multistep sampling is still preserved for better performance. DDDMs require no delicately designed samplers nor distillation on pre-trained distillation models. DDDMs train the diffusion model conditioned on an estimated target that was generated from previous training iterations of its own. To generate images, samples generated from the previous time step are also taken into consideration, guiding the generation process iteratively. We further propose Pseudo-LPIPS, a novel metric loss that is more robust to various values of hyperparameter. Despite its simplicity, the proposed approach can achieve strong performance in benchmark datasets. Our model achieves FID scores of 2.57 and 2.33 on CIFAR-10 in one-step and two-step sampling respectively, surpassing those obtained from GANs and distillation-based models. By extending the sampling to 1000 steps, we further reduce FID score to 1.79, aligning with state-of-the-art methods in the literature. For ImageNet 64x64, our approach stands as a competitive contender against leading models.
Authors: Chunhui Zhang, Li Liu, Hao Wen, Xi Zhou, Yanfeng Wang
Abstract: Multi-modal object tracking (MMOT) is an emerging field that combines data from various modalities, \eg vision (RGB), depth, thermal infrared, event, language and audio, to estimate the state of an arbitrary object in a video sequence. It is of great significance for many applications such as autonomous driving and intelligent surveillance. In recent years, MMOT has received more and more attention. However, existing MMOT algorithms mainly focus on two modalities (\eg RGB+depth, RGB+thermal infrared, and RGB+language). To leverage more modalities, some recent efforts have been made to learn a unified visual object tracking model for any modality. Additionally, some large-scale multi-modal tracking benchmarks have been established by simultaneously providing more than two modalities, such as vision-language-audio (\eg WebUAV-3M) and vision-depth-language (\eg UniMod1K). To track the latest progress in MMOT, we conduct a comprehensive investigation in this report. Specifically, we first divide existing MMOT tasks into five main categories, \ie RGBL tracking, RGBE tracking, RGBD tracking, RGBT tracking, and miscellaneous (RGB+X), where X can be any modality, such as language, depth, and event. Then, we analyze and summarize each MMOT task, focusing on widely used datasets and mainstream tracking algorithms based on their technical paradigms (\eg self-supervised learning, prompt learning, knowledge distillation, generative models, and state space models). Finally, we maintain a continuously updated paper list for MMOT at https://github.com/983632847/Awesome-Multimodal-Object-Tracking.
URLs: https://github.com/983632847/Awesome-Multimodal-Object-Tracking.
Authors: Fan Liu, Liang Yao, Chuanyi Zhang, Ting Wu, Xinlei Zhang, Xiruo Jiang, Jun Zhou
Abstract: Detecting objects from Unmanned Aerial Vehicles (UAV) is often hindered by a large number of small objects, resulting in low detection accuracy. To address this issue, mainstream approaches typically utilize multi-stage inferences. Despite their remarkable detecting accuracies, real-time efficiency is sacrificed, making them less practical to handle real applications. To this end, we propose to improve the single-stage inference accuracy through learning scale-invariant features. Specifically, a Scale-Invariant Feature Disentangling module is designed to disentangle scale-related and scale-invariant features. Then an Adversarial Feature Learning scheme is employed to enhance disentanglement. Finally, scale-invariant features are leveraged for robust UAV-based object detection. Furthermore, we construct a multi-modal UAV object detection dataset, State-Air, which incorporates annotated UAV state parameters. We apply our approach to three state-of-the-art lightweight detection frameworks on three benchmark datasets, including State-Air. Extensive experiments demonstrate that our approach can effectively improve model accuracy. Our code and dataset are provided in Supplementary Materials and will be publicly available once the paper is accepted.
Authors: Gu\'enol\'e Fiche, Simon Leglaive, Xavier Alameda-Pineda, Francesc Moreno-Noguer
Abstract: Human Mesh Recovery (HMR) from a single RGB image is a highly ambiguous problem, as similar 2D projections can correspond to multiple 3D interpretations. Nevertheless, most HMR methods overlook this ambiguity and make a single prediction without accounting for the associated uncertainty. A few approaches generate a distribution of human meshes, enabling the sampling of multiple predictions; however, none of them is competitive with the latest single-output model when making a single prediction. This work proposes a new approach based on masked generative modeling. By tokenizing the human pose and shape, we formulate the HMR task as generating a sequence of discrete tokens conditioned on an input image. We introduce MEGA, a MaskEd Generative Autoencoder trained to recover human meshes from images and partial human mesh token sequences. Given an image, our flexible generation scheme allows us to predict a single human mesh in deterministic mode or to generate multiple human meshes in stochastic mode. MEGA enables us to propose multiple outputs and to evaluate the uncertainty of the predictions. Experiments on in-the-wild benchmarks show that MEGA achieves state-of-the-art performance in deterministic and stochastic modes, outperforming single-output and multi-output approaches.
Authors: Hongyuan Dong, Jiawen Li, Bohong Wu, Jiacong Wang, Yuan Zhang, Haoyuan Guo
Abstract: Image captioning has long been regarded as a fundamental task in visual understanding. Recently, however, few large vision-language model (LVLM) research discusses model's image captioning performance because of the outdated short-caption benchmarks and unreliable evaluation metrics. In this work, we propose to benchmark detail image caption task by curating high-quality evaluation datasets annotated by human experts, GPT-4V and Gemini-1.5-Pro. We also design a more reliable caption evaluation metric called CAPTURE (CAPtion evaluation by exTracting and coUpling coRE information). CAPTURE extracts visual elements, e.g., objects, attributes and relations from captions, and then matches these elements through three stages, achieving the highest consistency with expert judgements over other rule-based or model-based caption metrics. The proposed benchmark and metric provide reliable evaluation for LVLM's detailed image captioning ability. Guided by this evaluation, we further explore to unleash LVLM's detail caption capabilities by synthesizing high-quality data through a five-stage data construction pipeline. Our pipeline only uses a given LVLM itself and other open-source tools, without any human or GPT-4V annotation in the loop. Experiments show that the proposed data construction strategy significantly improves model-generated detail caption data quality for LVLMs with leading performance, and the data quality can be further improved in a self-looping paradigm. All code and dataset will be publicly available at https://github.com/foundation-multimodal-models/CAPTURE.
URLs: https://github.com/foundation-multimodal-models/CAPTURE.
Authors: Wenchao Sun, Xuewu Lin, Yining Shi, Chuang Zhang, Haoran Wu, Sifa Zheng
Abstract: The well-established modular autonomous driving system is decoupled into different standalone tasks, e.g. perception, prediction and planning, suffering from information loss and error accumulation across modules. In contrast, end-to-end paradigms unify multi-tasks into a fully differentiable framework, allowing for optimization in a planning-oriented spirit. Despite the great potential of end-to-end paradigms, both the performance and efficiency of existing methods are not satisfactory, particularly in terms of planning safety. We attribute this to the computationally expensive BEV (bird's eye view) features and the straightforward design for prediction and planning. To this end, we explore the sparse representation and review the task design for end-to-end autonomous driving, proposing a new paradigm named SparseDrive. Concretely, SparseDrive consists of a symmetric sparse perception module and a parallel motion planner. The sparse perception module unifies detection, tracking and online mapping with a symmetric model architecture, learning a fully sparse representation of the driving scene. For motion prediction and planning, we review the great similarity between these two tasks, leading to a parallel design for motion planner. Based on this parallel design, which models planning as a multi-modal problem, we propose a hierarchical planning selection strategy , which incorporates a collision-aware rescore module, to select a rational and safe trajectory as the final planning output. With such effective designs, SparseDrive surpasses previous state-of-the-arts by a large margin in performance of all tasks, while achieving much higher training and inference efficiency. Code will be avaliable at https://github.com/swc-17/SparseDrive for facilitating future research.
Authors: Zixian Guo, Ming Liu, Zhilong Ji, Jinfeng Bai, Yiwen Guo, Wangmeng Zuo
Abstract: Learning a skill generally relies on both practical experience by doer and insightful high-level guidance by instructor. Will this strategy also work well for solving complex non-convex optimization problems? Here, a common gradient-based optimizer acts like a disciplined doer, making locally optimal update at each step. Recent methods utilize large language models (LLMs) to optimize solutions for concrete problems by inferring from natural language instructions, akin to a high-level instructor. In this paper, we show that these two optimizers are complementary to each other, suggesting a collaborative optimization approach. The gradient-based optimizer and LLM-based optimizer are combined in an interleaved manner. We instruct LLMs using task descriptions and timely optimization trajectories recorded during gradient-based optimization. Inferred results from LLMs are used as restarting points for the next stage of gradient optimization. By leveraging both the locally rigorous gradient-based optimizer and the high-level deductive LLM-based optimizer, our combined optimization method consistently yields improvements over competitive baseline prompt tuning methods. Our results demonstrate the synergistic effect of conventional gradient-based optimization and the inference ability of LLMs. The code is released at https://github.com/guozix/LLM-catalyst.
Authors: Wenxuan Liu, Sai Qian Zhang
Abstract: Diffusion Transformers (DiTs) have recently gained substantial attention in both industrial and academic fields for their superior visual generation capabilities, outperforming traditional diffusion models that use U-Net. However,the enhanced performance of DiTs also comes with high parameter counts and implementation costs, seriously restricting their use on resource-limited devices such as mobile phones. To address these challenges, we introduce the Hybrid Floating-point Quantization for DiT(HQ-DiT), an efficient post-training quantization method that utilizes 4-bit floating-point (FP) precision on both weights and activations for DiT inference. Compared to fixed-point quantization (e.g., INT8), FP quantization, complemented by our proposed clipping range selection mechanism, naturally aligns with the data distribution within DiT, resulting in a minimal quantization error. Furthermore, HQ-DiT also implements a universal identity mathematical transform to mitigate the serious quantization error caused by the outliers. The experimental results demonstrate that DiT can achieve extremely low-precision quantization (i.e., 4 bits) with negligible impact on performance. Our approach marks the first instance where both weights and activations in DiTs are quantized to just 4 bits, with only a 0.12 increase in sFID on ImageNet.
Authors: Masashi Hatano, Ryo Hachiuma, Ryo Fujii, Hideo Saito
Abstract: We address a novel cross-domain few-shot learning task (CD-FSL) with multimodal input and unlabeled target data for egocentric action recognition. This paper simultaneously tackles two critical challenges associated with egocentric action recognition in CD-FSL settings: (1) the extreme domain gap in egocentric videos (\eg, daily life vs. industrial domain) and (2) the computational cost for real-world applications. We propose MM-CDFSL, a domain-adaptive and computationally efficient approach designed to enhance adaptability to the target domain and improve inference speed. To address the first challenge, we propose the incorporation of multimodal distillation into the student RGB model using teacher models. Each teacher model is trained independently on source and target data for its respective modality. Leveraging only unlabeled target data during multimodal distillation enhances the student model's adaptability to the target domain. We further introduce ensemble masked inference, a technique that reduces the number of input tokens through masking. In this approach, ensemble prediction mitigates the performance degradation caused by masking, effectively addressing the second issue. Our approach outperformed the state-of-the-art CD-FSL approaches with a substantial margin on multiple egocentric datasets, improving by an average of 6.12/6.10 points for 1-shot/5-shot settings while achieving $2.2$ times faster inference speed. Project page: https://masashi-hatano.github.io/MM-CDFSL/
Authors: Honghao Fu, Yufei Wang, Wenhan Yang, Bihan Wen
Abstract: Image quality assessment (IQA) plays a critical role in selecting high-quality images and guiding compression and enhancement methods in a series of applications. The blind IQA, which assesses the quality of in-the-wild images containing complex authentic distortions without reference images, poses greater challenges. Existing methods are limited to modeling a uniform distribution with local patches and are bothered by the gap between low and high-level visions (caused by widely adopted pre-trained classification networks). In this paper, we propose a novel IQA method called diffusion priors-based IQA (DP-IQA), which leverages the prior knowledge from the pre-trained diffusion model with its excellent powers to bridge semantic gaps in the perception of the visual quality of images. Specifically, we use pre-trained stable diffusion as the backbone, extract multi-level features from the denoising U-Net during the upsampling process at a specified timestep, and decode them to estimate the image quality score. The text and image adapters are adopted to mitigate the domain gap for downstream tasks and correct the information loss caused by the variational autoencoder bottleneck. Finally, we distill the knowledge in the above model into a CNN-based student model, significantly reducing the parameter to enhance applicability, with the student model performing similarly or even better than the teacher model surprisingly. Experimental results demonstrate that our DP-IQA achieves state-of-the-art results on various in-the-wild datasets with better generalization capability, which shows the superiority of our method in global modeling and utilizing the hierarchical feature clues of diffusion for evaluating image quality.
Authors: Stavros Diolatzis, Tobias Zirr, Alexandr Kuznetsov, Georgios Kopanas, Anton Kaplanyan
Abstract: In the wake of many new ML-inspired approaches for reconstructing and representing high-quality 3D content, recent hybrid and explicitly learned representations exhibit promising performance and quality characteristics. However, their scaling to higher dimensions is challenging, e.g. when accounting for dynamic content with respect to additional parameters such as material properties, illumination, or time. In this paper, we tackle these challenges for an explicit representations based on Gaussian mixture models. With our solutions, we arrive at efficient fitting of compact N-dimensional Gaussian mixtures and enable efficient evaluation at render time: For fast fitting and evaluation, we introduce a high-dimensional culling scheme that efficiently bounds N-D Gaussians, inspired by Locality Sensitive Hashing. For adaptive refinement yet compact representation, we introduce a loss-adaptive density control scheme that incrementally guides the use of additional capacity towards missing details. With these tools we can for the first time represent complex appearance that depends on many input dimensions beyond position or viewing angle within a compact, explicit representation optimized in minutes and rendered in milliseconds.
Authors: Kai Wu, Boyuan Jiang, Zhengkai Jiang, Qingdong He, Donghao Luo, Shengzhi Wang, Qingwen Liu, Chengjie Wang
Abstract: Multimodal large language models (MLLMs) contribute a powerful mechanism to understanding visual information building on large language models. However, MLLMs are notorious for suffering from hallucinations, especially when generating lengthy, detailed descriptions for images. Our analysis reveals that hallucinations stem from the inherent summarization mechanism of large language models, leading to excessive dependence on linguistic tokens while neglecting vision information. In this paper, we propose NoiseBoost, a broadly applicable and simple method for alleviating hallucinations for MLLMs through the integration of noise feature perturbations. Noise perturbation acts as a regularizer, facilitating a balanced distribution of attention weights among visual and linguistic tokens. Despite its simplicity, NoiseBoost consistently enhances the performance of MLLMs across common training strategies, including supervised fine-tuning and reinforcement learning. Further, NoiseBoost pioneerly enables semi-supervised learning for MLLMs, unleashing the power of unlabeled data. Comprehensive experiments demonstrate that NoiseBoost improves dense caption accuracy by 8.1% with human evaluation and achieves comparable results with 50% of the data by mining unlabeled data. Code and models are available at https://kaiwu5.github.io/noiseboost.
Authors: Miriam Navarro, \'Alvaro Becerra, Roberto Daza, Ruth Cobos, Aythami Morales, Julian Fierrez
Abstract: In this paper, we present an approach in the Multimodal Learning Analytics field. Within this approach, we have developed a tool to visualize and analyze eye movement data collected during learning sessions in online courses. The tool is named VAAD (an acronym for Visual Attention Analysis Dashboard). These eye movement data have been gathered using an eye-tracker and subsequently processed and visualized for interpretation. The purpose of the tool is to conduct a descriptive analysis of the data by facilitating its visualization, enabling the identification of differences and learning patterns among various learner populations. Additionally, it integrates a predictive module capable of anticipating learner activities during a learning session. Consequently, VAAD holds the potential to offer valuable insights into online learning behaviors from both descriptive and predictive perspectives.
Authors: Jinrui Yang, Xianhang Li, Druv Pai, Yuyin Zhou, Yi Ma, Yaodong Yu, Cihang Xie
Abstract: CRATE, a white-box transformer architecture designed to learn compressed and sparse representations, offers an intriguing alternative to standard vision transformers (ViTs) due to its inherent mathematical interpretability. Despite extensive investigations into the scaling behaviors of language and vision transformers, the scalability of CRATE remains an open question which this paper aims to address. Specifically, we propose CRATE-$\alpha$, featuring strategic yet minimal modifications to the sparse coding block in the CRATE architecture design, and a light training recipe designed to improve the scalability of CRATE. Through extensive experiments, we demonstrate that CRATE-$\alpha$ can effectively scale with larger model sizes and datasets. For example, our CRATE-$\alpha$-B substantially outperforms the prior best CRATE-B model accuracy on ImageNet classification by 3.7%, achieving an accuracy of 83.2%. Meanwhile, when scaling further, our CRATE-$\alpha$-L obtains an ImageNet classification accuracy of 85.1%. More notably, these model performance improvements are achieved while preserving, and potentially even enhancing the interpretability of learned CRATE models, as we demonstrate through showing that the learned token representations of increasingly larger trained CRATE-$\alpha$ models yield increasingly higher-quality unsupervised object segmentation of images. The project page is https://rayjryang.github.io/CRATE-alpha/.
Authors: Jianghao Shen, Xue Nan, Tianfu Wu
Abstract: Learning 3D scene representation from a single-view image is a long-standing fundamental problem in computer vision, with the inherent ambiguity in predicting contents unseen from the input view. Built on the recently proposed 3D Gaussian Splatting (3DGS), the Splatter Image method has made promising progress on fast single-image novel view synthesis via learning a single 3D Gaussian for each pixel based on the U-Net feature map of an input image. However, it has limited expressive power to represent occluded components that are not observable in the input view. To address this problem, this paper presents a Hierarchical Splatter Image method in which a pixel is worth more than one 3D Gaussians. Specifically, each pixel is represented by a parent 3D Gaussian and a small number of child 3D Gaussians. Parent 3D Gaussians are learned as done in the vanilla Splatter Image. Child 3D Gaussians are learned via a lightweight Multi-Layer Perceptron (MLP) which takes as input the projected image features of a parent 3D Gaussian and the embedding of a target camera view. Both parent and child 3D Gaussians are learned end-to-end in a stage-wise way. The joint condition of input image features from eyes of the parent Gaussians and the target camera position facilitates learning to allocate child Gaussians to ``see the unseen'', recovering the occluded details that are often missed by parent Gaussians. In experiments, the proposed method is tested on the ShapeNet-SRN and CO3D datasets with state-of-the-art performance obtained, especially showing promising capabilities of reconstructing occluded contents in the input view.
Authors: Aditya Ganeshan, Ryan Y. Huang, Xianghao Xu, R. Kenny Jones, Daniel Ritchie
Abstract: The ability to edit 3D assets from natural language presents a compelling paradigm to aid in the democratization of 3D content creation. However, while natural language is often effective at communicating general intent, it is poorly suited for specifying precise manipulation. To address this gap, we introduce ParSEL, a system that enables controllable editing of high-quality 3D assets from natural language. Given a segmented 3D mesh and an editing request, ParSEL produces a parameterized editing program. Adjusting the program parameters allows users to explore shape variations with a precise control over the magnitudes of edits. To infer editing programs which align with an input edit request, we leverage the abilities of large-language models (LLMs). However, while we find that LLMs excel at identifying initial edit operations, they often fail to infer complete editing programs, and produce outputs that violate shape semantics. To overcome this issue, we introduce Analytical Edit Propagation (AEP), an algorithm which extends a seed edit with additional operations until a complete editing program has been formed. Unlike prior methods, AEP searches for analytical editing operations compatible with a range of possible user edits through the integration of computer algebra systems for geometric analysis. Experimentally we demonstrate ParSEL's effectiveness in enabling controllable editing of 3D objects through natural language requests over alternative system designs.
Authors: Dixuan Lin, Yuxiang Zhang, Mengcheng Li, Yebin Liu, Wei Jing, Qi Yan, Qianying Wang, Hongwen Zhang
Abstract: In this paper, we introduce 4DHands, a robust approach to recovering interactive hand meshes and their relative movement from monocular inputs. Our approach addresses two major limitations of previous methods: lacking a unified solution for handling various hand image inputs and neglecting the positional relationship of two hands within images. To overcome these challenges, we develop a transformer-based architecture with novel tokenization and feature fusion strategies. Specifically, we propose a Relation-aware Two-Hand Tokenization (RAT) method to embed positional relation information into the hand tokens. In this way, our network can handle both single-hand and two-hand inputs and explicitly leverage relative hand positions, facilitating the reconstruction of intricate hand interactions in real-world scenarios. As such tokenization indicates the relative relationship of two hands, it also supports more effective feature fusion. To this end, we further develop a Spatio-temporal Interaction Reasoning (SIR) module to fuse hand tokens in 4D with attention and decode them into 3D hand meshes and relative temporal movements. The efficacy of our approach is validated on several benchmark datasets. The results on in-the-wild videos and real-world scenarios demonstrate the superior performances of our approach for interactive hand reconstruction. More video results can be found on the project page: https://4dhands.github.io.
Authors: Jeffrey D. Rudie, Rachit Saluja, David A. Weiss, Pierre Nedelec, Evan Calabrese, John B. Colby, Benjamin Laguna, John Mongan, Steve Braunstein, Christopher P. Hess, Andreas M. Rauschecker, Leo P. Sugrue, Javier E. Villanueva-Meyer
Abstract: The University of California San Francisco Brain Metastases Stereotactic Radiosurgery (UCSF-BMSR) dataset is a public, clinical, multimodal brain MRI dataset consisting of 560 brain MRIs from 412 patients with expert annotations of 5136 brain metastases. Data consists of registered and skull stripped T1 post-contrast, T1 pre-contrast, FLAIR and subtraction (T1 pre-contrast - T1 post-contrast) images and voxelwise segmentations of enhancing brain metastases in NifTI format. The dataset also includes patient demographics, surgical status and primary cancer types. The UCSF-BSMR has been made publicly available in the hopes that researchers will use these data to push the boundaries of AI applications for brain metastases. The dataset is freely available for non-commercial use at https://imagingdatasets.ucsf.edu/dataset/1
Authors: Xue Yang, Zifeng Liu, Xiaohu Tang, Rongxing Lu, Bo Liu
Abstract: With the emergence of privacy leaks in federated learning, secure aggregation protocols that mainly adopt either homomorphic encryption or threshold secret sharing have been widely developed for federated learning to protect the privacy of the local training data of each client. However, these existing protocols suffer from many shortcomings, such as the dependence on a trusted third party, the vulnerability to clients being corrupted, low efficiency, the trade-off between security and fault tolerance, etc. To solve these disadvantages, we propose an efficient and multi-private key secure aggregation scheme for federated learning. Specifically, we skillfully modify the variant ElGamal encryption technique to achieve homomorphic addition operation, which has two important advantages: 1) The server and each client can freely select public and private keys without introducing a trust third party and 2) Compared to the variant ElGamal encryption, the plaintext space is relatively large, which is more suitable for the deep model. Besides, for the high dimensional deep model parameter, we introduce a super-increasing sequence to compress multi-dimensional data into 1-D, which can greatly reduce encryption and decryption times as well as communication for ciphertext transmission. Detailed security analyses show that our proposed scheme achieves the semantic security of both individual local gradients and the aggregated result while achieving optimal robustness in tolerating both client collusion and dropped clients. Extensive simulations demonstrate that the accuracy of our scheme is almost the same as the non-private approach, while the efficiency of our scheme is much better than the state-of-the-art homomorphic encryption-based secure aggregation schemes. More importantly, the efficiency advantages of our scheme will become increasingly prominent as the number of model parameters increases.
Authors: Noriyuki Kojima, Hadar Averbuch-Elor, Yoav Artzi
Abstract: Key to tasks that require reasoning about natural language in visual contexts is grounding words and phrases to image regions. However, observing this grounding in contemporary models is complex, even if it is generally expected to take place if the task is addressed in a way that is conductive to generalization. We propose a framework to jointly study task performance and phrase grounding, and propose three benchmarks to study the relation between the two. Our results show that contemporary models demonstrate inconsistency between their ability to ground phrases and solve tasks. We show how this can be addressed through brute-force training on ground phrasing annotations, and analyze the dynamics it creates. Code and at available at https://github.com/lil-lab/phrase_grounding.
Authors: Megh Shukla, Mathieu Salzmann, Alexandre Alahi
Abstract: Deep heteroscedastic regression involves jointly optimizing the mean and covariance of the predicted distribution using the negative log-likelihood. However, recent works show that this may result in sub-optimal convergence due to the challenges associated with covariance estimation. While the literature addresses this by proposing alternate formulations to mitigate the impact of the predicted covariance, we focus on improving the predicted covariance itself. We study two questions: (1) Does the predicted covariance truly capture the randomness of the predicted mean? (2) In the absence of supervision, how can we quantify the accuracy of covariance estimation? We address (1) with a Taylor Induced Covariance (TIC), which captures the randomness of the predicted mean by incorporating its gradient and curvature through the second order Taylor polynomial. Furthermore, we tackle (2) by introducing a Task Agnostic Correlations (TAC) metric, which combines the notion of correlations and absolute error to evaluate the covariance. We evaluate TIC-TAC across multiple experiments spanning synthetic and real-world datasets. Our results show that not only does TIC accurately learn the covariance, it additionally facilitates an improved convergence of the negative log-likelihood. Our code is available at https://github.com/vita-epfl/TIC-TAC
Authors: Richard Osuala, Smriti Joshi, Apostolia Tsirikoglou, Lidia Garrucho, Walter H. L. Pinaya, Oliver Diaz, Karim Lekadir
Abstract: Despite its benefits for tumour detection and treatment, the administration of contrast agents in dynamic contrast-enhanced MRI (DCE-MRI) is associated with a range of issues, including their invasiveness, bioaccumulation, and a risk of nephrogenic systemic fibrosis. This study explores the feasibility of producing synthetic contrast enhancements by translating pre-contrast T1-weighted fat-saturated breast MRI to their corresponding first DCE-MRI sequence leveraging the capabilities of a generative adversarial network (GAN). Additionally, we introduce a Scaled Aggregate Measure (SAMe) designed for quantitatively evaluating the quality of synthetic data in a principled manner and serving as a basis for selecting the optimal generative model. We assess the generated DCE-MRI data using quantitative image quality metrics and apply them to the downstream task of 3D breast tumour segmentation. Our results highlight the potential of post-contrast DCE-MRI synthesis in enhancing the robustness of breast tumour segmentation models via data augmentation. Our code is available at https://github.com/RichardObi/pre_post_synthesis.
Authors: Simisola Odimayo, Chollette C. Olisah, Khadija Mohammed
Abstract: Alzheimer's disease (AD), the predominant form of dementia, is a growing global challenge, emphasizing the urgent need for accurate and early diagnosis. Current clinical diagnoses rely on radiologist expert interpretation, which is prone to human error. Deep learning has thus far shown promise for early AD diagnosis. However, existing methods often overlook focal structural atrophy critical for enhanced understanding of the cerebral cortex neurodegeneration. This paper proposes a deep learning framework that includes a novel structure-focused neurodegeneration CNN architecture named SNeurodCNN and an image brightness enhancement preprocessor using gamma correction. The SNeurodCNN architecture takes as input the focal structural atrophy features resulting from segmentation of brain structures captured through magnetic resonance imaging (MRI). As a result, the architecture considers only necessary CNN components, which comprises of two downsampling convolutional blocks and two fully connected layers, for achieving the desired classification task, and utilises regularisation techniques to regularise learnable parameters. Leveraging mid-sagittal and para-sagittal brain image viewpoints from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, our framework demonstrated exceptional performance. The para-sagittal viewpoint achieved 97.8% accuracy, 97.0% specificity, and 98.5% sensitivity, while the mid-sagittal viewpoint offered deeper insights with 98.1% accuracy, 97.2% specificity, and 99.0% sensitivity. Model analysis revealed the ability of SNeurodCNN to capture the structural dynamics of mild cognitive impairment (MCI) and AD in the frontal lobe, occipital lobe, cerebellum, temporal, and parietal lobe, suggesting its potential as a brain structural change digi-biomarker for early AD diagnosis. This work can be reproduced using code we made available on GitHub.
Authors: Haibo Jin, Ruoxi Chen, Andy Zhou, Yang Zhang, Haohan Wang
Abstract: The discovery of "jailbreaks" to bypass safety filters of Large Language Models (LLMs) and harmful responses have encouraged the community to implement safety measures. One major safety measure is to proactively test the LLMs with jailbreaks prior to the release. Therefore, such testing will require a method that can generate jailbreaks massively and efficiently. In this paper, we follow a novel yet intuitive strategy to generate jailbreaks in the style of the human generation. We propose a role-playing system that assigns four different roles to the user LLMs to collaborate on new jailbreaks. Furthermore, we collect existing jailbreaks and split them into different independent characteristics using clustering frequency and semantic patterns sentence by sentence. We organize these characteristics into a knowledge graph, making them more accessible and easier to retrieve. Our system of different roles will leverage this knowledge graph to generate new jailbreaks, which have proved effective in inducing LLMs to generate unethical or guideline-violating responses. In addition, we also pioneer a setting in our system that will automatically follow the government-issued guidelines to generate jailbreaks to test whether LLMs follow the guidelines accordingly. We refer to our system as GUARD (Guideline Upholding through Adaptive Role-play Diagnostics). We have empirically validated the effectiveness of GUARD on three cutting-edge open-sourced LLMs (Vicuna-13B, LongChat-7B, and Llama-2-7B), as well as a widely-utilized commercial LLM (ChatGPT). Moreover, our work extends to the realm of vision language models (MiniGPT-v2 and Gemini Vision Pro), showcasing GUARD's versatility and contributing valuable insights for the development of safer, more reliable LLM-based applications across diverse modalities.
Authors: Lei Li, Tianfang Zhang, Xinglin Zhang, Jiaqi Liu, Bingqi Ma, Yan Luo, Tao Chen
Abstract: Within the domain of medical analysis, extensive research has explored the potential of mutual learning between Masked Autoencoders(MAEs) and multimodal data. However, the impact of MAEs on intermodality remains a key challenge. We introduce MedFLIP, a Fast Language-Image Pre-training method for Medical analysis. We explore MAEs for zero-shot learning with crossed domains, which enhances the model's ability to learn from limited data, a common scenario in medical diagnostics. We verify that masking an image does not affect inter-modal learning. Furthermore, we propose the SVD loss to enhance the representation learning for characteristics of medical images, aiming to improve classification accuracy by leveraging the structural intricacies of such data. Our theory posits that masking encourages semantic preservation, robust feature extraction, regularization, domain adaptation, and invariance learning. Lastly, we validate using language will improve the zero-shot performance for the medical image analysis. MedFLIP's scaling of the masking process marks an advancement in the field, offering a pathway to rapid and precise medical image analysis without the traditional computational bottlenecks. Through experiments and validation, MedFLIP demonstrates efficient performance improvements, helps for future research and application in medical diagnostics.
Authors: Jiwoong Im, Nayoung Kwon, Taewoo Park, Jiheon Woo, Jaeho Lee, Yongjune Kim
Abstract: We propose a communication-efficient collaborative inference framework in the domain of edge inference, focusing on the efficient use of vision transformer (ViT) models. The partitioning strategy of conventional collaborative inference fails to reduce communication cost because of the inherent architecture of ViTs maintaining consistent layer dimensions across the entire transformer encoder. Therefore, instead of employing the partitioning strategy, our framework utilizes a lightweight ViT model on the edge device, with the server deploying a complicated ViT model. To enhance communication efficiency and achieve the classification accuracy of the server model, we propose two strategies: 1) attention-aware patch selection and 2) entropy-aware image transmission. Attention-aware patch selection leverages the attention scores generated by the edge device's transformer encoder to identify and select the image patches critical for classification. This strategy enables the edge device to transmit only the essential patches to the server, significantly improving communication efficiency. Entropy-aware image transmission uses min-entropy as a metric to accurately determine whether to depend on the lightweight model on the edge device or to request the inference from the server model. In our framework, the lightweight ViT model on the edge device acts as a semantic encoder, efficiently identifying and selecting the crucial image information required for the classification task. Our experiments demonstrate that the proposed collaborative inference framework can reduce communication overhead by 68% with only a minimal loss in accuracy compared to the server model on the ImageNet dataset.
Authors: Weijian Sun, Yanbo Jia, Qi Zeng, Zihao Liu, Jiang Liao, Yue Li, Xianfeng Li
Abstract: Deep-learning-based techniques have been widely adopted for autonomous driving software stacks for mass production in recent years, focusing primarily on perception modules, with some work extending this method to prediction modules. However, the downstream planning and control modules are still designed with hefty handcrafted rules, dominated by optimization-based methods such as quadratic programming or model predictive control. This results in a performance bottleneck for autonomous driving systems in that corner cases simply cannot be solved by enumerating hand-crafted rules. We present a deep-learning-based approach that brings prediction, decision, and planning modules together with the attempt to overcome the rule-based methods' deficiency in real-world applications of autonomous driving, especially for urban scenes. The DNN model we proposed is solely trained with 10 hours of human driver data, and it supports all mass-production ADAS features available on the market to date. This method is deployed onto a Jiyue test car with no modification to its factory-ready sensor set and compute platform. the feasibility, usability, and commercial potential are demonstrated in this article.
Authors: Qiyue Wang, Wu Xue, Xiaoke Zhang, Fang Jin, James Hahn
Abstract: It is critically important to detect the content of liver fat as it is related to cardiac complications and cardiovascular disease mortality. However, existing methods are either associated with high cost and/or medical complications (e.g., liver biopsy, imaging technology) or only roughly estimate the grades of steatosis. In this paper, we propose a deep neural network to estimate the percentage of liver fat using only body shapes. The proposed is composed of a flexible baseline network and a lightweight Attention module. The attention module is trained to generate discriminative and diverse features which significant improve the performance. In order to validate the method, we perform extensive tests on the public medical dataset. The results verify that our proposed method yields state-of-the-art performance with Root mean squared error (RMSE) of 5.26% and R-Squared value over 0.8. It offers an accurate and more accessible assessment of hepatic steatosis.
Authors: Yiyang Zhou, Zhiyuan Fan, Dongjie Cheng, Sihan Yang, Zhaorun Chen, Chenhang Cui, Xiyao Wang, Yun Li, Linjun Zhang, Huaxiu Yao
Abstract: Large Vision-Language Models (LVLMs) have made substantial progress by integrating pre-trained large language models (LLMs) and vision models through instruction tuning. Despite these advancements, LVLMs often exhibit the hallucination phenomenon, where generated text responses appear linguistically plausible but contradict the input image, indicating a misalignment between image and text pairs. This misalignment arises because the model tends to prioritize textual information over visual input, even when both the language model and visual representations are of high quality. Existing methods leverage additional models or human annotations to curate preference data and enhance modality alignment through preference optimization. These approaches may not effectively reflect the target LVLM's preferences, making the curated preferences easily distinguishable. Our work addresses these challenges by proposing the Calibrated Self-Rewarding (CSR) approach, which enables the model to self-improve by iteratively generating candidate responses, evaluating the reward for each response, and curating preference data for fine-tuning. In the reward modeling, we employ a step-wise strategy and incorporate visual constraints into the self-rewarding process to place greater emphasis on visual input. Empirical results demonstrate that CSR enhances performance and reduces hallucinations across ten benchmarks and tasks, achieving substantial improvements over existing methods by 7.62%. Our empirical results are further supported by rigorous theoretical analysis, under mild assumptions, verifying the effectiveness of introducing visual constraints into the self-rewarding paradigm. Additionally, CSR shows compatibility with different vision-language models and the ability to incrementally improve performance through iterative fine-tuning. Our data and code are available at https://github.com/YiyangZhou/CSR.
Authors: Nicklas Hansen, Jyothir S V, Vlad Sobal, Yann LeCun, Xiaolong Wang, Hao Su
Abstract: Whole-body control for humanoids is challenging due to the high-dimensional nature of the problem, coupled with the inherent instability of a bipedal morphology. Learning from visual observations further exacerbates this difficulty. In this work, we explore highly data-driven approaches to visual whole-body humanoid control based on reinforcement learning, without any simplifying assumptions, reward design, or skill primitives. Specifically, we propose a hierarchical world model in which a high-level agent generates commands based on visual observations for a low-level agent to execute, both of which are trained with rewards. Our approach produces highly performant control policies in 8 tasks with a simulated 56-DoF humanoid, while synthesizing motions that are broadly preferred by humans. Code and videos: https://nicklashansen.com/rlpuppeteer
Authors: Rui-Jie Zhu, Ziqing Wang, Leilani Gilpin, Jason K. Eshraghian
Abstract: Autonomous driving demands an integrated approach that encompasses perception, prediction, and planning, all while operating under strict energy constraints to enhance scalability and environmental sustainability. We present Spiking Autonomous Driving (SAD), the first unified Spiking Neural Network (SNN) to address the energy challenges faced by autonomous driving systems through its event-driven and energy-efficient nature. SAD is trained end-to-end and consists of three main modules: perception, which processes inputs from multi-view cameras to construct a spatiotemporal bird's eye view; prediction, which utilizes a novel dual-pathway with spiking neurons to forecast future states; and planning, which generates safe trajectories considering predicted occupancy, traffic rules, and ride comfort. Evaluated on the nuScenes dataset, SAD achieves competitive performance in perception, prediction, and planning tasks, while drawing upon the energy efficiency of SNNs. This work highlights the potential of neuromorphic computing to be applied to energy-efficient autonomous driving, a critical step toward sustainable and safety-critical automotive technology. Our code is available at \url{https://github.com/ridgerchu/SAD}.
Authors: Matthew Watson, Divyashree Shivakumar Sreepathihalli, Francois Chollet, Martin Gorner, Kiranbir Sodhia, Ramesh Sampath, Tirth Patel, Haifeng Jin, Neel Kovelamudi, Gabriel Rasskin, Samaneh Saadat, Luke Wood, Chen Qian, Jonathan Bischof, Ian Stenbit, Abheesht Sharma, Anshuman Mishra
Abstract: We present the Keras domain packages KerasCV and KerasNLP, extensions of the Keras API for Computer Vision and Natural Language Processing workflows, capable of running on either JAX, TensorFlow, or PyTorch. These domain packages are designed to enable fast experimentation, with a focus on ease-of-use and performance. We adopt a modular, layered design: at the library's lowest level of abstraction, we provide building blocks for creating models and data preprocessing pipelines, and at the library's highest level of abstraction, we provide pretrained ``task" models for popular architectures such as Stable Diffusion, YOLOv8, GPT2, BERT, Mistral, CLIP, Gemma, T5, etc. Task models have built-in preprocessing, pretrained weights, and can be fine-tuned on raw inputs. To enable efficient training, we support XLA compilation for all models, and run all preprocessing via a compiled graph of TensorFlow operations using the tf.data API. The libraries are fully open-source (Apache 2.0 license) and available on GitHub.