Authors: Chunhui Zhang, Yawen Cui, Weilin Lin, Guanjie Huang, Yan Rong, Li Liu, Shiguang Shan
Abstract: The recent wave of foundation models has witnessed tremendous success in computer vision (CV) and beyond, with the segment anything model (SAM) having sparked a passion for exploring task-agnostic visual foundation models. Empowered by its remarkable zero-shot generalization, SAM is currently challenging numerous traditional paradigms in CV, delivering extraordinary performance not only in various image segmentation and multi-modal segmentation (\eg, text-to-mask) tasks, but also in the video domain. Additionally, the latest released SAM 2 is once again sparking research enthusiasm in the realm of promptable visual segmentation for both images and videos. However, existing surveys mainly focus on SAM in various image processing tasks, a comprehensive and in-depth review in the video domain is notably absent. To address this gap, this work conducts a systematic review on SAM for videos in the era of foundation models. As the first to review the progress of SAM for videos, this work focuses on its applications to various tasks by discussing its recent advances, and innovation opportunities of developing foundation models on broad applications. We begin with a brief introduction to the background of SAM and video-related research domains. Subsequently, we present a systematic taxonomy that categorizes existing methods into three key areas: video understanding, video generation, and video editing, analyzing and summarizing their advantages and limitations. Furthermore, comparative results of SAM-based and current state-of-the-art methods on representative benchmarks, as well as insightful analysis are offered. Finally, we discuss the challenges faced by current research and envision several future research directions in the field of SAM for video and beyond.
Authors: Zongze Wu, Nicholas Kolkin, Jonathan Brandt, Richard Zhang, Eli Shechtman
Abstract: We address the challenges of precise image inversion and disentangled image editing in the context of few-step diffusion models. We introduce an encoder based iterative inversion technique. The inversion network is conditioned on the input image and the reconstructed image from the previous step, allowing for correction of the next reconstruction towards the input image. We demonstrate that disentangled controls can be easily achieved in the few-step diffusion model by conditioning on an (automatically generated) detailed text prompt. To manipulate the inverted image, we freeze the noise maps and modify one attribute in the text prompt (either manually or via instruction based editing driven by an LLM), resulting in the generation of a new image similar to the input image with only one attribute changed. It can further control the editing strength and accept instructive text prompt. Our approach facilitates realistic text-guided image edits in real-time, requiring only 8 number of functional evaluations (NFEs) in inversion (one-time cost) and 4 NFEs per edit. Our method is not only fast, but also significantly outperforms state-of-the-art multi-step diffusion editing techniques.
Authors: Dongshuo Yin, Leiyi Hu, Bin Li, Youqun Zhang, Xue Yang
Abstract: Pre-training & fine-tuning can enhance the transferring efficiency and performance in visual tasks. Recent delta-tuning methods provide more options for visual classification tasks. Despite their success, existing visual delta-tuning art fails to exceed the upper limit of full fine-tuning on challenging tasks like object detection and segmentation. To find a competitive alternative to full fine-tuning, we propose the Multi-cognitive Visual Adapter (Mona) tuning, a novel adapter-based tuning method. First, we introduce multiple vision-friendly filters into the adapter to enhance its ability to process visual signals, while previous methods mainly rely on language-friendly linear filters. Second, we add the scaled normalization layer in the adapter to regulate the distribution of input features for visual filters. To fully demonstrate the practicality and generality of Mona, we conduct experiments on multiple representative visual tasks, including instance segmentation on COCO, semantic segmentation on ADE20K, object detection on Pascal VOC, oriented object detection on DOTA/STAR, and image classification on three common datasets. Exciting results illustrate that Mona surpasses full fine-tuning on all these tasks, and is the only delta-tuning method outperforming full fine-tuning on the above various tasks. For example, Mona achieves 1% performance gain on the COCO dataset compared to full fine-tuning. Comprehensive results suggest that Mona-tuning is more suitable for retaining and utilizing the capabilities of pre-trained models than full fine-tuning. We will make the code publicly available.
Authors: Nicholas Konz, Maciej A. Mazurowski
Abstract: In recent years, there has been interest in how geometric properties such as intrinsic dimension (ID) of a neural network's hidden representations evolve through its layers, and how such properties are predictive of important model behavior such as generalization ability. However, evidence has begun to emerge that such behavior can change significantly depending on the domain of the network's training data, such as natural versus medical images. Here, we further this inquiry by exploring how the ID of a network's learned representations evolves through its layers, in essence, characterizing how the network successively refines the information content of input data to be used for predictions. Analyzing eleven natural and medical image datasets across six network architectures, we find that the shape of this ID evolution curve differs noticeably between natural and medical image models: medical image models peak in representation ID earlier in the network, implying a difference in the image features and their abstractness that are typically used for downstream tasks in these domains. Additionally, we discover a strong correlation of this peak representation ID with the ID of the data in its input space, implying that the intrinsic information content of a model's learned representations is guided by that of the data it was trained on. Overall, our findings emphasize notable discrepancies in network behavior between natural and non-natural imaging domains regarding hidden representation information content, and provide further insights into how a network's learned features are shaped by its training data.
Authors: Daniele Rege Cambrin, Gabriele Scaffidi Militone, Luca Colomba, Giovanni Malnati, Daniele Apiletti, Paolo Garza
Abstract: Designing effective game tutorials is crucial for a smooth learning curve for new players, especially in games with many rules and complex core mechanics. Evaluating the effectiveness of these tutorials usually requires multiple iterations with testers who have no prior knowledge of the game. Recent Vision-Language Models (VLMs) have demonstrated significant capabilities in understanding and interpreting visual content. VLMs can analyze images, provide detailed insights, and answer questions about their content. They can recognize objects, actions, and contexts in visual data, making them valuable tools for various applications, including automated game testing. In this work, we propose an automated game-testing solution to evaluate the quality of game tutorials. Our approach leverages VLMs to analyze frames from video game tutorials, answer relevant questions to simulate human perception, and provide feedback. This feedback is compared with expected results to identify confusing or problematic scenes and highlight potential errors for developers. In addition, we publish complete tutorial videos and annotated frames from different game versions used in our tests. This solution reduces the need for extensive manual testing, especially by speeding up and simplifying the initial development stages of the tutorial to improve the final game experience.
Authors: Yabin Wang, Zhiwu Huang, Su Zhou, Adam Prugel-Bennett, Xiaopeng Hong
Abstract: The diffusion of deepfake technologies has sparked serious concerns about its potential misuse across various domains, prompting the urgent need for robust detection methods. Despite advancement, many current approaches prioritize short-term gains at expense of long-term effectiveness. This paper critiques the overly specialized approach of fine-tuning pre-trained models solely with a penny-wise objective on a single deepfake dataset, while disregarding the pound-wise balance for generalization and knowledge retention. To address this "Penny-Wise and Pound-Foolish" issue, we propose a novel learning framework (PoundNet) for generalization of deepfake detection on a pre-trained vision-language model. PoundNet incorporates a learnable prompt design and a balanced objective to preserve broad knowledge from upstream tasks (object classification) while enhancing generalization for downstream tasks (deepfake detection). We train PoundNet on a standard single deepfake dataset, following common practice in the literature. We then evaluate its performance across 10 public large-scale deepfake datasets with 5 main evaluation metrics-forming the largest benchmark test set for assessing the generalization ability of deepfake detection models, to our knowledge. The comprehensive benchmark evaluation demonstrates the proposed PoundNet is significantly less "Penny-Wise and Pound-Foolish", achieving a remarkable improvement of 19% in deepfake detection performance compared to state-of-the-art methods, while maintaining a strong performance of 63% on object classification tasks, where other deepfake detection models tend to be ineffective. Code and data are open-sourced at https://github.com/iamwangyabin/PoundNet.
Authors: Kshitij Bhardwaj
Abstract: While Vision Transformers (ViTs) are extremely effective at computer vision tasks and are replacing convolutional neural networks as the new state-of-the-art, they are complex and memory-intensive models. In order to effectively run these models on resource-constrained mobile/edge systems, there is a need to not only compress these models but also to optimize them and convert them into deployment-friendly formats. To this end, this paper presents a combined pruning and quantization tool, called PQV-Mobile, to optimize vision transformers for mobile applications. The tool is able to support different types of structured pruning based on magnitude importance, Taylor importance, and Hessian importance. It also supports quantization from FP32 to FP16 and int8, targeting different mobile hardware backends. We demonstrate the capabilities of our tool and show important latency-memory-accuracy trade-offs for different amounts of pruning and int8 quantization with Facebook Data Efficient Image Transformer (DeiT) models. Our results show that even pruning a DeiT model by 9.375% and quantizing it to int8 from FP32 followed by optimizing for mobile applications, we find a latency reduction by 7.18X with a small accuracy loss of 2.24%. The tool is open source.
Authors: Nassim Ait Ali Braham, Conrad M Albrecht, Julien Mairal, Jocelyn Chanussot, Yi Wang, Xiao Xiang Zhu
Abstract: Foundation models have triggered a paradigm shift in computer vision and are increasingly being adopted in remote sensing, particularly for multispectral imagery. Yet, their potential in hyperspectral imaging (HSI) remains untapped due to the absence of comprehensive and globally representative hyperspectral datasets. To close this gap, we introduce SpectralEarth, a large-scale multi-temporal dataset designed to pretrain hyperspectral foundation models leveraging data from the Environmental Mapping and Analysis Program (EnMAP). SpectralEarth comprises 538,974 image patches covering 415,153 unique locations from more than 11,636 globally distributed EnMAP scenes spanning two years of archive. Additionally, 17.5% of these locations include multiple timestamps, enabling multi-temporal HSI analysis. Utilizing state-of-the-art self-supervised learning (SSL) algorithms, we pretrain a series of foundation models on SpectralEarth. We integrate a spectral adapter into classical vision backbones to accommodate the unique characteristics of HSI. In tandem, we construct four downstream datasets for land-cover and crop-type mapping, providing benchmarks for model evaluation. Experimental results support the versatility of our models, showcasing their generalizability across different tasks and sensors. We also highlight computational efficiency during model fine-tuning. The dataset, models, and source code will be made publicly available.
Authors: Zohaib Khan, Muhammad Khaquan, Omer Tafveez, Agha Ali Raza
Abstract: The Transformer architecture has revolutionized deep learning through its Self-Attention mechanism, which effectively captures contextual information. However, the memory footprint of Self-Attention presents significant challenges for long-sequence tasks. Grouped Query Attention (GQA) addresses this issue by grouping queries and mean-pooling the corresponding key-value heads - reducing the number of overall parameters and memory requirements in a flexible manner without adversely compromising model accuracy. In this work, we introduce enhancements to GQA, focusing on two novel approaches that deviate from the static nature of grouping: Key-Distributed GQA (KDGQA) and Dynamic Key-Distributed GQA (DGQA), which leverage information from the norms of the key heads to inform query allocation. Specifically, KDGQA looks at the ratios of the norms of the key heads during each forward pass, while DGQA examines the ratios of the norms as they evolve through training. Additionally, we present Perturbed GQA (PGQA) as a case-study, which introduces variability in (static) group formation via subtracting noise from the attention maps. Our experiments with up-trained Vision Transformers, for Image Classification on datasets such as CIFAR-10, CIFAR-100, Food101, and Tiny ImageNet, demonstrate the promise of these variants in improving upon the original GQA through more informed and adaptive grouping mechanisms: specifically ViT-L experiences accuracy gains of up to 8% when utilizing DGQA in comparison to GQA and other variants. We further analyze the impact of the number of Key-Value Heads on performance, underscoring the importance of utilizing query-key affinities.
Authors: Jihun Park, Jongmin Gim, Kyoungmin Lee, Seunghun Lee, Sunghoon Im
Abstract: We present Text-driven Object-Centric Style Transfer (TEXTOC), a novel method that guides style transfer at an object-centric level using textual inputs. The core of TEXTOC is our Patch-wise Co-Directional (PCD) loss, meticulously designed for precise object-centric transformations that are closely aligned with the input text. This loss combines a patch directional loss for text-guided style direction and a patch distribution consistency loss for even CLIP embedding distribution across object regions. It ensures a seamless and harmonious style transfer across object regions. Key to our method are the Text-Matched Patch Selection (TMPS) and Pre-fixed Region Selection (PRS) modules for identifying object locations via text, eliminating the need for segmentation masks. Lastly, we introduce an Adaptive Background Preservation (ABP) loss to maintain the original style and structural essence of the image's background. This loss is applied to dynamically identified background areas. Extensive experiments underline the effectiveness of our approach in creating visually coherent and textually aligned style transfers.
Authors: Mohammadreza Samadi, Fred X. Han, Mohammad Salameh, Hao Wu, Fengyu Sun, Chunhua Zhou, Di Niu
Abstract: Diffusion models have demonstrated strong performance in generative tasks, making them ideal candidates for image editing. Recent studies highlight their ability to apply desired edits effectively by following textual instructions, yet two key challenges persist. First, these models struggle to apply multiple edits simultaneously, resulting in computational inefficiencies due to their reliance on sequential processing. Second, relying on textual prompts to determine the editing region can lead to unintended alterations in other parts of the image. In this work, we introduce FunEditor, an efficient diffusion model designed to learn atomic editing functions and perform complex edits by aggregating simpler functions. This approach enables complex editing tasks, such as object movement, by aggregating multiple functions and applying them simultaneously to specific areas. FunEditor is 5 to 24 times faster inference than existing methods on complex tasks like object movement. Our experiments demonstrate that FunEditor significantly outperforms recent baselines, including both inference-time optimization methods and fine-tuned models, across various metrics, such as image quality assessment (IQA) and object-background consistency.
Authors: Shihan Peng, Hanyu Zhou, Hao Dong, Zhiwei Shi, Haoyue Liu, Yuxing Duan, Yi Chang, Luxin Yan
Abstract: Conventional frame camera is the mainstream sensor of the autonomous driving scene perception, while it is limited in adverse conditions, such as low light. Event camera with high dynamic range has been applied in assisting frame camera for the multimodal fusion, which relies heavily on the pixel-level spatial alignment between various modalities. Typically, existing multimodal datasets mainly place event and frame cameras in parallel and directly align them spatially via warping operation. However, this parallel strategy is less effective for multimodal fusion, since the large disparity exacerbates spatial misalignment due to the large event-frame baseline. We argue that baseline minimization can reduce alignment error between event and frame cameras. In this work, we introduce hybrid coaxial event-frame devices to build the multimodal system, and propose a coaxial stereo event camera (CoSEC) dataset for autonomous driving. As for the multimodal system, we first utilize the microcontroller to achieve time synchronization, and then spatially calibrate different sensors, where we perform intra- and inter-calibration of stereo coaxial devices. As for the multimodal dataset, we filter LiDAR point clouds to generate depth and optical flow labels using reference depth, which is further improved by fusing aligned event and frame data in nighttime conditions. With the help of the coaxial device, the proposed dataset can promote the all-day pixel-level multimodal fusion. Moreover, we also conduct experiments to demonstrate that the proposed dataset can improve the performance and generalization of the multimodal fusion.
Authors: Hefei Mei, Minjing Dong, Chang Xu
Abstract: Diffusion models (DMs) have demonstrated great potential in the field of adversarial robustness, where DM-based defense methods can achieve superior defense capability without adversarial training. However, they all require huge computational costs due to the usage of large-scale pre-trained DMs, making it difficult to conduct full evaluation under strong attacks and compare with traditional CNN-based methods. Simply reducing the network size and timesteps in DMs could significantly harm the image generation quality, which invalidates previous frameworks. To alleviate this issue, we redesign the diffusion framework from generating high-quality images to predicting distinguishable image labels. Specifically, we employ an image translation framework to learn many-to-one mapping from input samples to designed orthogonal image labels. Based on this framework, we introduce an efficient Image-to-Image diffusion classifier with a pruned U-Net structure and reduced diffusion timesteps. Besides the framework, we redesign the optimization objective of DMs to fit the target of image classification, where a new classification loss is incorporated in the DM-based image translation framework to distinguish the generated label from those of other classes. We conduct sufficient evaluations of the proposed classifier under various attacks on popular benchmarks. Extensive experiments show that our method achieves better adversarial robustness with fewer computational costs than DM-based and CNN-based methods. The code is available at https://github.com/hfmei/IDC.
Authors: Xiaoyue Mi, Fan Tang, Juan Cao, Peng Li, Yang Liu
Abstract: Personalized concept generation by tuning diffusion models with a few images raises potential legal and ethical concerns regarding privacy and intellectual property rights. Researchers attempt to prevent malicious personalization using adversarial perturbations. However, previous efforts have mainly focused on the effectiveness of protection while neglecting the visibility of perturbations. They utilize global adversarial perturbations, which introduce noticeable alterations to original images and significantly degrade visual quality. In this work, we propose the Visual-Friendly Concept Protection (VCPro) framework, which prioritizes the protection of key concepts chosen by the image owner through adversarial perturbations with lower perceptibility. To ensure these perturbations are as inconspicuous as possible, we introduce a relaxed optimization objective to identify the least perceptible yet effective adversarial perturbations, solved using the Lagrangian multiplier method. Qualitative and quantitative experiments validate that VCPro achieves a better trade-off between the visibility of perturbations and protection effectiveness, effectively prioritizing the protection of target concepts in images with less perceptible perturbations.
Authors: Kang Du, Zhihao Liang, Zeyu Wang
Abstract: We present GS-ID, a novel framework for illumination decomposition on Gaussian Splatting, achieving photorealistic novel view synthesis and intuitive light editing. Illumination decomposition is an ill-posed problem facing three main challenges: 1) priors for geometry and material are often lacking; 2) complex illumination conditions involve multiple unknown light sources; and 3) calculating surface shading with numerous light sources is computationally expensive. To address these challenges, we first introduce intrinsic diffusion priors to estimate the attributes for physically based rendering. Then we divide the illumination into environmental and direct components for joint optimization. Last, we employ deferred rendering to reduce the computational load. Our framework uses a learnable environment map and Spherical Gaussians (SGs) to represent light sources parametrically, therefore enabling controllable and photorealistic relighting on Gaussian Splatting. Extensive experiments and applications demonstrate that GS-ID produces state-of-the-art illumination decomposition results while achieving better geometry reconstruction and rendering performance.
Authors: Li Pan, Yupei Zhang, Qiushi Yang, Tan Li, Xiaohan Xing, Maximus C. F. Yeung, Zhen Chen
Abstract: Recently, multimodal deep learning, which integrates histopathology slides and molecular biomarkers, has achieved a promising performance in glioma grading. Despite great progress, due to the intra-modality complexity and inter-modality heterogeneity, existing studies suffer from inadequate histopathology representation learning and inefficient molecular-pathology knowledge alignment. These two issues hinder existing methods to precisely interpret diagnostic molecular-pathology features, thereby limiting their grading performance. Moreover, the real-world applicability of existing multimodal approaches is significantly restricted as molecular biomarkers are not always available during clinical deployment. To address these problems, we introduce a novel Focus on Focus (FoF) framework with paired pathology-genomic training and applicable pathology-only inference, enhancing molecular-pathology representation effectively. Specifically, we propose a Focus-oriented Representation Learning (FRL) module to encourage the model to identify regions positively or negatively related to glioma grading and guide it to focus on the diagnostic areas with a consistency constraint. To effectively link the molecular biomarkers to morphological features, we propose a Multi-view Cross-modal Alignment (MCA) module that projects histopathology representations into molecular subspaces, aligning morphological features with corresponding molecular biomarker status by supervised contrastive learning. Experiments on the TCGA GBM-LGG dataset demonstrate that our FoF framework significantly improves the glioma grading. Remarkably, our FoF achieves superior performance using only histopathology slides compared to existing multimodal methods. The source code is available at https://github.com/peterlipan/FoF.
Authors: Kouki Horio, Kiyoshi Nishikawa, Hitoshi Kiya
Abstract: We propose a novel method for privacy-preserving fine-tuning vision transformers (ViTs) with encrypted images. Conventional methods using encrypted images degrade model performance compared with that of using plain images due to the influence of image encryption. In contrast, the proposed encryption method using restricted random permutation matrices can provide a higher performance than the conventional ones.
Authors: Hongqiu Wang, Wei Wang, Haipeng Zhou, Huihui Xu, Shaozhi Wu, Lei Zhu
Abstract: Traditional shadow detectors often identify all shadow regions of static images or video sequences. This work presents the Referring Video Shadow Detection (RVSD), which is an innovative task that rejuvenates the classic paradigm by facilitating the segmentation of particular shadows in videos based on descriptive natural language prompts. This novel RVSD not only achieves segmentation of arbitrary shadow areas of interest based on descriptions (flexibility) but also allows users to interact with visual content more directly and naturally by using natural language prompts (interactivity), paving the way for abundant applications ranging from advanced video editing to virtual reality experiences. To pioneer the RVSD research, we curated a well-annotated RVSD dataset, which encompasses 86 videos and a rich set of 15,011 paired textual descriptions with corresponding shadows. To the best of our knowledge, this dataset is the first one for addressing RVSD. Based on this dataset, we propose a Referring Shadow-Track Memory Network (RSM-Net) for addressing the RVSD task. In our RSM-Net, we devise a Twin-Track Synergistic Memory (TSM) to store intra-clip memory features and hierarchical inter-clip memory features, and then pass these memory features into a memory read module to refine features of the current video frame for referring shadow detection. We also develop a Mixed-Prior Shadow Attention (MSA) to utilize physical priors to obtain a coarse shadow map for learning more visual features by weighting it with the input video frame. Experimental results show that our RSM-Net achieves state-of-the-art performance for RVSD with a notable Overall IOU increase of 4.4\%. Our code and dataset are available at https://github.com/whq-xxh/RVSD.
Authors: Wengang Zhou, Weichao Zhao, Hezhen Hu, Zecheng Li, Houqiang Li
Abstract: Sign language serves as the primary meaning of communication for the deaf-mute community. Different from spoken language, it commonly conveys information by the collaboration of manual features, i.e., hand gestures and body movements, and non-manual features, i.e., facial expressions and mouth cues. To facilitate communication between the deaf-mute and hearing people, a series of sign language understanding (SLU) tasks have been studied in recent years, including isolated/continuous sign language recognition (ISLR/CSLR), gloss-free sign language translation (GF-SLT) and sign language retrieval (SL-RT). Sign language recognition and translation aims to understand the semantic meaning conveyed by sign languages from gloss-level and sentence-level, respectively. In contrast, SL-RT focuses on retrieving sign videos or corresponding texts from a closed-set under the query-by-example search paradigm. These tasks investigate sign language topics from diverse perspectives and raise challenges in learning effective representation of sign language videos. To advance the development of sign language understanding, exploring a generalized model that is applicable across various SLU tasks is a profound research direction.
Authors: Umang Sharma, Jungkyu Park, Laura Heacock, Sumit Chopra, Krzysztof Geras
Abstract: Full Field Digital Mammograms (FFDMs) and Digital Breast Tomosynthesis (DBT) are the two most widely used imaging modalities for breast cancer screening. Although DBT has increased cancer detection compared to FFDM, its widespread adoption in clinical practice has been slowed by increased interpretation times and a perceived decrease in the conspicuity of specific lesion types. Specifically, the non-inferiority of DBT for microcalcifications remains under debate. Due to concerns about the decrease in visual acuity, combined DBT-FFDM acquisitions remain popular, leading to overall increased exam times and radiation dosage. Enabling DBT to provide diagnostic information present in both FFDM and DBT would reduce reliance on FFDM, resulting in a reduction in both quantities. We propose a machine learning methodology that learns high-level representations leveraging the complementary diagnostic signal from both DBT and FFDM. Experiments on a large-scale data set validate our claims and show that our representations enable more accurate breast lesion detection than any DBT- or FFDM-based model.
Authors: Yujia Gu, Xinyu Fang, Xueyuan Deng
Abstract: This study mainly introduces a method combining the Stable Diffusion Model (SDM) and Parameter-Efficient Fine-Tuning method for generating Chinese Landscape Paintings. This training process is accelerated by combining LoRA with pre-trained SDM and DreamBooth with pre-trained SDM, respectively. On the Chinese Landscape Paintings Internet dataset used in this paper, this study finds that SDM combined with DreamBooth exhibits superior performance, outperforming other models, including the generic pre-trained SDM and LoRA-based fine-tuning SDM. The SDM combined with DreamBooth achieves a FID of 12.75 on the dataset and outperforms all other models in terms of expert evaluation, highlighting the model's versatility in the field of Chinese Landscape Paintings given the unique identifier, high fidelity and high quality. This study illustrates the potential of specialised fine-tuning method to improve the performance of SDM on domain-specific tasks, particularly in the domain of Landscape Paintings.
Authors: Zhangquan Chen, Puhua Jiang, Ruqi Huang
Abstract: In this paper, we propose a novel learning-based framework for non-rigid point cloud matching, which can be trained purely on point clouds without any correspondence annotation but also be extended naturally to partial-to-full matching. Our key insight is to incorporate semantic features derived from large vision models (LVMs) to geometry-based shape feature learning. Our framework effectively leverages the structural information contained in the semantic features to address ambiguities arise from self-similarities among local geometries. Furthermore, our framework also enjoys the strong generalizability and robustness regarding partial observations of LVMs, leading to improvements in the regarding point cloud matching tasks. In order to achieve the above, we propose a pixel-to-point feature aggregation module, a local and global attention network as well as a geometrical similarity loss function. Experimental results show that our method achieves state-of-the-art results in matching non-rigid point clouds in both near-isometric and heterogeneous shape collection as well as more realistic partial and noisy data.
Authors: Jun Zhou, Chunsheng Liu, Faliang Chang, Wenqian Wang, Penghui Hao, Yiming Huang, Zhiqiang Yang
Abstract: Associating driver attention with driving scene across two fields of views (FOVs) is a hard cross-domain perception problem, which requires comprehensive consideration of cross-view mapping, dynamic driving scene analysis, and driver status tracking. Previous methods typically focus on a single view or map attention to the scene via estimated gaze, failing to exploit the implicit connection between them. Moreover, simple fusion modules are insufficient for modeling the complex relationships between the two views, making information integration challenging. To address these issues, we propose a novel method for end-to-end scene-associated driver attention estimation, called EraW-Net. This method enhances the most discriminative dynamic cues, refines feature representations, and facilitates semantically aligned cross-domain integration through a W-shaped architecture, termed W-Net. Specifically, a Dynamic Adaptive Filter Module (DAF-Module) is proposed to address the challenges of frequently changing driving environments by extracting vital regions. It suppresses the indiscriminately recorded dynamics and highlights crucial ones by innovative joint frequency-spatial analysis, enhancing the model's ability to parse complex dynamics. Additionally, to track driver states during non-fixed facial poses, we propose a Global Context Sharing Module (GCS-Module) to construct refined feature representations by capturing hierarchical features that adapt to various scales of head and eye movements. Finally, W-Net achieves systematic cross-view information integration through its "Encoding-Independent Partial Decoding-Fusion Decoding" structure, addressing semantic misalignment in heterogeneous data integration. Experiments demonstrate that the proposed method robustly and accurately estimates the mapping of driver attention in scene on large public datasets.
Authors: Jinming Liu, Yuntao Wei, Junyan Lin, Shengyang Zhao, Heming Sun, Zhibo Chen, Wenjun Zeng, Xin Jin
Abstract: We present a new image compression paradigm to achieve ``intelligently coding for machine'' by cleverly leveraging the common sense of Large Multimodal Models (LMMs). We are motivated by the evidence that large language/multimodal models are powerful general-purpose semantics predictors for understanding the real world. Different from traditional image compression typically optimized for human eyes, the image coding for machines (ICM) framework we focus on requires the compressed bitstream to more comply with different downstream intelligent analysis tasks. To this end, we employ LMM to \textcolor{red}{tell codec what to compress}: 1) first utilize the powerful semantic understanding capability of LMMs w.r.t object grounding, identification, and importance ranking via prompts, to disentangle image content before compression, 2) and then based on these semantic priors we accordingly encode and transmit objects of the image in order with a structured bitstream. In this way, diverse vision benchmarks including image classification, object detection, instance segmentation, etc., can be well supported with such a semantically structured bitstream. We dub our method ``\textit{SDComp}'' for ``\textit{S}emantically \textit{D}isentangled \textit{Comp}ression'', and compare it with state-of-the-art codecs on a wide variety of different vision tasks. SDComp codec leads to more flexible reconstruction results, promised decoded visual quality, and a more generic/satisfactory intelligent task-supporting ability.
Authors: Linghao Zheng, Xinyang Pu, Feng Xu
Abstract: The Segment Anything Model (SAM), a foundational model designed for promptable segmentation tasks, demonstrates exceptional generalization capabilities, making it highly promising for natural scene image segmentation. However, SAM's lack of pretraining on massive remote sensing images and its interactive structure limit its automatic mask prediction capabilities. In this paper, a Multi-Cognitive SAM-Based Instance Segmentation Model (MC-SAM SEG) is introduced to employ SAM on remote sensing domain. The SAM-Mona encoder utilizing the Multi-cognitive Visual Adapter (Mona) is conducted to facilitate SAM's transfer learning in remote sensing applications. The proposed method named MC-SAM SEG extracts high-quality features by fine-tuning the SAM-Mona encoder along with a feature aggregator. Subsequently, a pixel decoder and transformer decoder are designed for prompt-free mask generation and instance classification. The comprehensive experiments are conducted on the HRSID and WHU datasets for instance segmentation tasks on Synthetic Aperture Radar (SAR) images and optical remote sensing images respectively. The evaluation results indicate the proposed method surpasses other deep learning algorithms and verify its effectiveness and generalization.
Authors: Jianhua Zhu, Wenqi Zhao, Yu Li, Xingjian Hu, Liangcai Gao
Abstract: Handwritten Mathematical Expression Recognition (HMER) has extensive applications in automated grading and office automation. However, existing sequence-based decoding methods, which directly predict $\LaTeX$ sequences, struggle to understand and model the inherent tree structure of $\LaTeX$ and often fail to ensure syntactic correctness in the decoded results. To address these challenges, we propose a novel model named TAMER (Tree-Aware Transformer) for handwritten mathematical expression recognition. TAMER introduces an innovative Tree-aware Module while maintaining the flexibility and efficient training of Transformer. TAMER combines the advantages of both sequence decoding and tree decoding models by jointly optimizing sequence prediction and tree structure prediction tasks, which enhances the model's understanding and generalization of complex mathematical expression structures. During inference, TAMER employs a Tree Structure Prediction Scoring Mechanism to improve the structural validity of the generated $\LaTeX$ sequences. Experimental results on CROHME datasets demonstrate that TAMER outperforms traditional sequence decoding and tree decoding models, especially in handling complex mathematical structures, achieving state-of-the-art (SOTA) performance.
Authors: Tri Ton, Ji Woo Hong, SooHwan Eom, Jun Yeop Shim, Junyeong Kim, Chang D. Yoo
Abstract: Open-vocabulary 3D instance segmentation transcends traditional closed-vocabulary methods by enabling the identification of both previously seen and unseen objects in real-world scenarios. It leverages a dual-modality approach, utilizing both 3D point clouds and 2D multi-view images to generate class-agnostic object mask proposals. Previous efforts predominantly focused on enhancing 3D mask proposal models; consequently, the information that could come from 2D association to 3D was not fully exploited. This bias towards 3D data, while effective for familiar indoor objects, limits the system's adaptability to new and varied object types, where 2D models offer greater utility. Addressing this gap, we introduce Zero-Shot Dual-Path Integration Framework that equally values the contributions of both 3D and 2D modalities. Our framework comprises three components: 3D pathway, 2D pathway, and Dual-Path Integration. 3D pathway generates spatially accurate class-agnostic mask proposals of common indoor objects from 3D point cloud data using a pre-trained 3D model, while 2D pathway utilizes pre-trained open-vocabulary instance segmentation model to identify a diverse array of object proposals from multi-view RGB-D images. In Dual-Path Integration, our Conditional Integration process, which operates in two stages, filters and merges the proposals from both pathways adaptively. This process harmonizes output proposals to enhance segmentation capabilities. Our framework, utilizing pre-trained models in a zero-shot manner, is model-agnostic and demonstrates superior performance on both seen and unseen data, as evidenced by comprehensive evaluations on the ScanNet200 and qualitative results on ARKitScenes datasets.
Authors: Zunjie Xiao, Xiaoqing Zhang, Risa Higashita, Jiang Liu
Abstract: Ophthalmic image segmentation serves as a critical foundation for ocular disease diagnosis. Although fully convolutional neural networks (CNNs) are commonly employed for segmentation, they are constrained by inductive biases and face challenges in establishing long-range dependencies. Transformer-based models address these limitations but introduce substantial computational overhead. Recently, a simple yet efficient Multilayer Perceptron (MLP) architecture was proposed for image classification, achieving competitive performance relative to advanced transformers. However, its effectiveness for ophthalmic image segmentation remains unexplored. In this paper, we introduce MM-UNet, an efficient Mixed MLP model tailored for ophthalmic image segmentation. Within MM-UNet, we propose a multi-scale MLP (MMLP) module that facilitates the interaction of features at various depths through a grouping strategy, enabling simultaneous capture of global and local information. We conducted extensive experiments on both a private anterior segment optical coherence tomography (AS-OCT) image dataset and a public fundus image dataset. The results demonstrated the superiority of our MM-UNet model in comparison to state-of-the-art deep segmentation networks.
Authors: Xiangyu Chen, Yihao Liu, Yuandong Pu, Wenlong Zhang, Jiantao Zhou, Yu Qiao, Chao Dong
Abstract: Building a unified model for general low-level vision tasks holds significant research and practical value. Current methods encounter several critical issues. Multi-task restoration approaches can address multiple degradation-to-clean restoration tasks, while their applicability to tasks with different target domains (e.g., image stylization) is limited. Methods like PromptGIP can handle multiple input-target domains but rely on the Masked Autoencoder (MAE) paradigm. Consequently, they are tied to the ViT architecture, resulting in suboptimal image reconstruction quality. In addition, these methods are sensitive to prompt image content and often struggle with low-frequency information processing. In this paper, we propose a Visual task Prompt-based Image Processing (VPIP) framework to overcome these challenges. VPIP employs visual task prompts to manage tasks with different input-target domains and allows flexible selection of backbone network suitable for general tasks. Besides, a new prompt cross-attention is introduced to facilitate interaction between the input and prompt information. Based on the VPIP framework, we train a low-level vision generalist model, namely GenLV, on 30 diverse tasks. Experimental results show that GenLV can successfully address a variety of low-level tasks, significantly outperforming existing methods both quantitatively and qualitatively. Codes are available at https://github.com/chxy95/GenLV.
Authors: Xihua Sheng, Li Li, Dong Liu, Shiqi Wang
Abstract: Deep video compression has made remarkable process in recent years, with the majority of advancements concentrated on P-frame coding. Although efforts to enhance B-frame coding are ongoing, their compression performance is still far behind that of traditional bi-directional video codecs. In this paper, we introduce a bi-directional deep contextual video compression scheme tailored for B-frames, termed DCVC-B, to improve the compression performance of deep B-frame coding. Our scheme mainly has three key innovations. First, we develop a bi-directional motion difference context propagation method for effective motion difference coding, which significantly reduces the bit cost of bi-directional motions. Second, we propose a bi-directional contextual compression model and a corresponding bi-directional temporal entropy model, to make better use of the multi-scale temporal contexts. Third, we propose a hierarchical quality structure-based training strategy, leading to an effective bit allocation across large groups of pictures (GOP). Experimental results show that our DCVC-B achieves an average reduction of 26.6% in BD-Rate compared to the reference software for H.265/HEVC under random access conditions. Remarkably, it surpasses the performance of the H.266/VVC reference software on certain test datasets under the same configuration.
Authors: Duo Su, Junjie Hou, Guang Li, Ren Togo, Rui Song, Takahiro Ogawa, Miki Haseyama
Abstract: This paper presents our method for the generative track of The First Dataset Distillation Challenge at ECCV 2024. Since the diffusion model has become the mainstay of generative models because of its high-quality generative effects, we focus on distillation methods based on the diffusion model. Considering that the track can only generate a fixed number of images in 10 minutes using a generative model for CIFAR-100 and Tiny-ImageNet datasets, we need to use a generative model that can generate images at high speed. In this study, we proposed a novel generative dataset distillation method based on Stable Diffusion. Specifically, we use the SDXL-Turbo model which can generate images at high speed and quality. Compared to other diffusion models that can only generate images per class (IPC) = 1, our method can achieve an IPC = 10 for Tiny-ImageNet and an IPC = 20 for CIFAR-100, respectively. Additionally, to generate high-quality distilled datasets for CIFAR-100 and Tiny-ImageNet, we use the class information as text prompts and post data augmentation for the SDXL-Turbo model. Experimental results show the effectiveness of the proposed method, and we achieved third place in the generative track of the ECCV 2024 DD Challenge. Codes are available at https://github.com/Guang000/BANKO.
Authors: Xingyue Lin, Xingjian Hu, Shuai Peng, Jianhua Zhu, Liangcai Gao
Abstract: Sketch, a powerful artistic technique to capture essential visual information about real-world objects, is increasingly gaining attention in the image synthesis field. However, evaluating the quality of synthesized sketches presents unique unsolved challenges. Current evaluation methods for sketch synthesis are inadequate due to the lack of a unified benchmark dataset, over-reliance on classification accuracy for recognizability, and unfair evaluation of sketches with different levels of simplification. To address these issues, we introduce SketchRef, a benchmark dataset comprising 4 categories of reference photos--animals, human faces, human bodies, and common objects--alongside novel evaluation metrics. Considering that classification accuracy is insufficient to measure the structural consistency between a sketch and its reference photo, we propose the mean Object Keypoint Similarity (mOKS) metric, utilizing pose estimation to assess structure-level recognizability. To ensure fair evaluation sketches with different simplification levels, we propose a recognizability calculation method constrained by simplicity. We also collect 8K responses from art enthusiasts, validating the effectiveness of our proposed evaluation methods. We hope this work can provide a comprehensive evaluation of sketch synthesis algorithms, thereby aligning their performance more closely with human understanding.
Authors: Sayan Kumar Chaki, Zeynep Sonat Baltaci, Elliot Vincent, Remi Emonet, Fabienne Vial-Bonacci, Christelle Bahier-Porte, Mathieu Aubry, Thierry Fournel
Abstract: This paper aims to develop the study of historical printed ornaments with modern unsupervised computer vision. We highlight three complex tasks that are of critical interest to book historians: clustering, element discovery, and unsupervised change localization. For each of these tasks, we introduce an evaluation benchmark, and we adapt and evaluate state-of-the-art models. Our Rey's Ornaments dataset is designed to be a representative example of a set of ornaments historians would be interested in. It focuses on an XVIIIth century bookseller, Marc-Michel Rey, providing a consistent set of ornaments with a wide diversity and representative challenges. Our results highlight the limitations of state-of-the-art models when faced with real data and show simple baselines such as k-means or congealing can outperform more sophisticated approaches on such data. Our dataset and code can be found at https://printed-ornaments.github.io/.
Authors: Kai Li, Jingbo Chen, Yupeng Deng, Yu Meng, Diyou Liu, Junxian Ma, Chenhao Wang
Abstract: Building Footprint Extraction (BFE) in off-nadir aerial images often relies on roof segmentation and roof-to-footprint offset prediction, then drugging roof-to-footprint via the offset. However, the results from this multi-stage inference are not applicable in data production, because of the low quality of masks given by prediction. To solve this problem, we proposed OBMv2 in this paper, which supports both end-to-end and promptable polygonal footprint prediction. Different from OBM, OBMv2 using a newly proposed Self Offset Attention (SOFA) to bridge the performance gap on bungalow and skyscraper, which realized a real end-to-end footprint polygon prediction without postprocessing. %, such as Non-Maximum Suppression (NMS) and Distance NMS (DNMS). % To fully use information contained in roof masks, building masks and offsets, we proposed a Multi-level Information SyStem (MISS) for footprint prediction, with which OBMv2 can predict footprints even with insufficient predictions. Additionally, to squeeze information from the same model, we were inspired by Retrieval-Augmented Generation (RAG) in Nature Language Processing and proposed "RAG in BFE" problem. To verify the effectiveness of the proposed method, experiments were conducted on open datasets BONAI and OmniCity-view3. A generalization test was also conducted on Huizhou test set. The code will be available at \url{https://github.com/likaiucas/OBM}.
Authors: Xin Di, Long Peng, Peizhe Xia, Wenbo Li, Renjing Pei, Yang Cao, Yang Wang, Zheng-Jun Zha
Abstract: Burst super-resolution aims to reconstruct high-resolution images with higher quality and richer details by fusing the sub-pixel information from multiple burst low-resolution frames. In BusrtSR, the key challenge lies in extracting the base frame's content complementary sub-pixel details while simultaneously suppressing high-frequency noise disturbance. Existing methods attempt to extract sub-pixels by modeling inter-frame relationships frame by frame while overlooking the mutual correlations among multi-current frames and neglecting the intra-frame interactions, leading to inaccurate and noisy sub-pixels for base frame super-resolution. Further, existing methods mainly employ static upsampling with fixed parameters to improve spatial resolution for all scenes, failing to perceive the sub-pixel distribution difference across multiple frames and cannot balance the fusion weights of different frames, resulting in over-smoothed details and artifacts. To address these limitations, we introduce a novel Query Mamba Burst Super-Resolution (QMambaBSR) network, which incorporates a Query State Space Model (QSSM) and Adaptive Up-sampling module (AdaUp). Specifically, based on the observation that sub-pixels have consistent spatial distribution while random noise is inconsistently distributed, a novel QSSM is proposed to efficiently extract sub-pixels through inter-frame querying and intra-frame scanning while mitigating noise interference in a single step. Moreover, AdaUp is designed to dynamically adjust the upsampling kernel based on the spatial distribution of multi-frame sub-pixel information in the different burst scenes, thereby facilitating the reconstruction of the spatial arrangement of high-resolution details. Extensive experiments on four popular synthetic and real-world benchmarks demonstrate that our method achieves a new state-of-the-art performance.
Authors: Alessio Devoto, Federico Alvetreti, Jary Pomponi, Paolo Di Lorenzo, Pasquale Minervini, Simone Scardapane
Abstract: Recently, foundation models based on Vision Transformers (ViTs) have become widely available. However, their fine-tuning process is highly resource-intensive, and it hinders their adoption in several edge or low-energy applications. To this end, in this paper we introduce an efficient fine-tuning method for ViTs called $\textbf{ALaST}$ ($\textit{Adaptive Layer Selection Fine-Tuning for Vision Transformers}$) to speed up the fine-tuning process while reducing computational cost, memory load, and training time. Our approach is based on the observation that not all layers are equally critical during fine-tuning, and their importance varies depending on the current mini-batch. Therefore, at each fine-tuning step, we adaptively estimate the importance of all layers and we assign what we call ``compute budgets'' accordingly. Layers that were allocated lower budgets are either trained with a reduced number of input tokens or kept frozen. Freezing a layer reduces the computational cost and memory usage by preventing updates to its weights, while discarding tokens removes redundant data, speeding up processing and reducing memory requirements. We show that this adaptive compute allocation enables a nearly-optimal schedule for distributing computational resources across layers, resulting in substantial reductions in training time (up to 1.5x), FLOPs (up to 2x), and memory load (up to 2x) compared to traditional full fine-tuning approaches. Additionally, it can be successfully combined with other parameter-efficient fine-tuning methods, such as LoRA.
Authors: Martin Hermann Paul Fuchs, Behnood Rasti, Beg\"um Demir
Abstract: The development of learning-based hyperspectral image (HSI) compression models has recently attracted significant interest. Existing models predominantly utilize convolutional filters, which capture only local dependencies. Furthermore, they often incur high training costs and exhibit substantial computational complexity. To address these limitations, in this paper we propose Hyperspectral Compression Transformer (HyCoT) that is a transformer-based autoencoder for pixelwise HSI compression. Additionally, we introduce an efficient training strategy to accelerate the training process. Experimental results on the HySpecNet-11k dataset demonstrate that HyCoT surpasses the state-of-the-art across various compression ratios by over 1 dB with significantly reduced computational requirements. Our code and pre-trained weights are publicly available at https://git.tu-berlin.de/rsim/hycot .
Authors: Miaoge Li, Jingcai Guo, Richard Yi Da Xu, Dongsheng Wang, Xiaofeng Cao, Song Guo
Abstract: Compositional Zero-Shot Learning (CZSL) aims to recognize novel \textit{state-object} compositions by leveraging the shared knowledge of their primitive components. Despite considerable progress, effectively calibrating the bias between semantically similar multimodal representations, as well as generalizing pre-trained knowledge to novel compositional contexts, remains an enduring challenge. In this paper, our interest is to revisit the conditional transport (CT) theory and its homology to the visual-semantics interaction in CZSL and further, propose a novel Trisets Consistency Alignment framework (dubbed TsCA) that well-addresses these issues. Concretely, we utilize three distinct yet semantically homologous sets, i.e., patches, primitives, and compositions, to construct pairwise CT costs to minimize their semantic discrepancies. To further ensure the consistency transfer within these sets, we implement a cycle-consistency constraint that refines the learning by guaranteeing the feature consistency of the self-mapping during transport flow, regardless of modality. Moreover, we extend the CT plans to an open-world setting, which enables the model to effectively filter out unfeasible pairs, thereby speeding up the inference as well as increasing the accuracy. Extensive experiments are conducted to verify the effectiveness of the proposed method.
Authors: Yang Nan, Huichi Zhou, Xiaodan Xing, Guang Yang
Abstract: Recent advancements in Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities across diverse tasks, garnering significant attention in AI communities. However, their performance and reliability in specialized domains such as medicine remain insufficiently assessed. In particular, most assessments over-concentrate in evaluating VLMs based on simple Visual Question Answering (VQA) on multi-modality data, while ignoring the in-depth characteristic of LVLMs. In this study, we introduce RadVUQA, a novel Radiological Visual Understanding and Question Answering benchmark, to comprehensively evaluate existing LVLMs. RadVUQA mainly validates LVLMs across five dimensions: 1) Anatomical understanding, assessing the models' ability to visually identify biological structures; 2) Multimodal comprehension, which involves the capability of interpreting linguistic and visual instructions to produce desired outcomes; 3) Quantitative and spatial reasoning, evaluating the models' spatial awareness and proficiency in combining quantitative analysis with visual and linguistic information; 4) Physiological knowledge, measuring the models' capability to comprehend functions and mechanisms of organs and systems; and 5) Robustness, which assesses the models' capabilities against unharmonised and synthetic data. The results indicate that both generalized LVLMs and medical-specific LVLMs have critical deficiencies with weak multimodal comprehension and quantitative reasoning capabilities. Our findings reveal the large gap between existing LVLMs and clinicians, highlighting the urgent need for more robust and intelligent LVLMs. The code and dataset will be available after the acceptance of this paper.
Authors: Kaixiang Yang, Wenqi Shan, Xudong Li, Xuan Wang, Xikai Yang, Xi Wang, Pheng-Ann Heng, Qiang Li, Zhiwei Wang
Abstract: Multi-modal brain tumor segmentation typically involves four magnetic resonance imaging (MRI) modalities, while incomplete modalities significantly degrade performance. Existing solutions employ explicit or implicit modality adaptation, aligning features across modalities or learning a fused feature robust to modality incompleteness. They share a common goal of encouraging each modality to express both itself and the others. However, the two expression abilities are entangled as a whole in a seamless feature space, resulting in prohibitive learning burdens. In this paper, we propose DeMoSeg to enhance the modality adaptation by Decoupling the task of representing the ego and other Modalities for robust incomplete multi-modal Segmentation. The decoupling is super lightweight by simply using two convolutions to map each modality onto four feature sub-spaces. The first sub-space expresses itself (Self-feature), while the remaining sub-spaces substitute for other modalities (Mutual-features). The Self- and Mutual-features interactively guide each other through a carefully-designed Channel-wised Sparse Self-Attention (CSSA). After that, a Radiologist-mimic Cross-modality expression Relationships (RCR) is introduced to have available modalities provide Self-feature and also `lend' their Mutual-features to compensate for the absent ones by exploiting the clinical prior knowledge. The benchmark results on BraTS2020, BraTS2018 and BraTS2015 verify the DeMoSeg's superiority thanks to the alleviated modality adaptation difficulty. Concretely, for BraTS2020, DeMoSeg increases Dice by at least 0.92%, 2.95% and 4.95% on whole tumor, tumor core and enhanced tumor regions, respectively, compared to other state-of-the-arts. Codes are at https://github.com/kk42yy/DeMoSeg
Authors: Wei Sun, Xiaosong Zhang, Fang Wan, Yanzhao Zhou, Yuan Li, Qixiang Ye, Jianbin Jiao
Abstract: Novel View Synthesis (NVS) without Structure-from-Motion (SfM) pre-processed camera poses--referred to as SfM-free methods--is crucial for promoting rapid response capabilities and enhancing robustness against variable operating conditions. Recent SfM-free methods have integrated pose optimization, designing end-to-end frameworks for joint camera pose estimation and NVS. However, most existing works rely on per-pixel image loss functions, such as L2 loss. In SfM-free methods, inaccurate initial poses lead to misalignment issue, which, under the constraints of per-pixel image loss functions, results in excessive gradients, causing unstable optimization and poor convergence for NVS. In this study, we propose a correspondence-guided SfM-free 3D Gaussian splatting for NVS. We use correspondences between the target and the rendered result to achieve better pixel alignment, facilitating the optimization of relative poses between frames. We then apply the learned poses to optimize the entire scene. Each 2D screen-space pixel is associated with its corresponding 3D Gaussians through approximated surface rendering to facilitate gradient back propagation. Experimental results underline the superior performance and time efficiency of the proposed approach compared to the state-of-the-art baselines.
Authors: Tianyi Xu, Yiji Zhou, Xiaotao Hu, Kai Zhang, Anran Zhang, Xingye Qiu, Jun Xu
Abstract: Arbitrary-scale super-resolution (ASSR) aims to learn a single model for image super-resolution at arbitrary magnifying scales. Existing ASSR networks typically comprise an off-the-shelf scale-agnostic feature extractor and an arbitrary scale upsampler. These feature extractors often use fixed network architectures to address different ASSR inference tasks, each of which is characterized by an input image and an upsampling scale. However, this overlooks the difficulty variance of super-resolution on different inference scenarios, where simple images or small SR scales could be resolved with less computational effort than difficult images or large SR scales. To tackle this difficulty variability, in this paper, we propose a Task-Aware Dynamic Transformer (TADT) as an input-adaptive feature extractor for efficient image ASSR. Our TADT consists of a multi-scale feature extraction backbone built upon groups of Multi-Scale Transformer Blocks (MSTBs) and a Task-Aware Routing Controller (TARC). The TARC predicts the inference paths within feature extraction backbone, specifically selecting MSTBs based on the input images and SR scales. The prediction of inference path is guided by a new loss function to trade-off the SR accuracy and efficiency. Experiments demonstrate that, when working with three popular arbitrary-scale upsamplers, our TADT achieves state-of-the-art ASSR performance when compared with mainstream feature extractors, but with relatively fewer computational costs. The code will be publicly released.
Authors: Sanchayan Vivekananthan
Abstract: This paper examines three major generative modelling frameworks: Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Stable Diffusion models. VAEs are effective at learning latent representations but frequently yield blurry results. GANs can generate realistic images but face issues such as mode collapse. Stable Diffusion models, while producing high-quality images with strong semantic coherence, are demanding in terms of computational resources. Additionally, the paper explores how incorporating Grounding DINO and Grounded SAM with Stable Diffusion improves image accuracy by utilising sophisticated segmentation and inpainting techniques. The analysis guides on selecting suitable models for various applications and highlights areas for further research.
Authors: Xiangdong Zhang, Shaofeng Zhang, Junchi Yan
Abstract: Masked autoencoder has been widely explored in point cloud self-supervised learning, whereby the point cloud is generally divided into visible and masked parts. These methods typically include an encoder accepting visible patches (normalized) and corresponding patch centers (position) as input, with the decoder accepting the output of the encoder and the centers (position) of the masked parts to reconstruct each point in the masked patches. Then, the pre-trained encoders are used for downstream tasks. In this paper, we show a motivating empirical result that when directly feeding the centers of masked patches to the decoder without information from the encoder, it still reconstructs well. In other words, the centers of patches are important and the reconstruction objective does not necessarily rely on representations of the encoder, thus preventing the encoder from learning semantic representations. Based on this key observation, we propose a simple yet effective method, i.e., learning to Predict Centers for Point Masked AutoEncoders (PCP-MAE) which guides the model to learn to predict the significant centers and use the predicted centers to replace the directly provided centers. Specifically, we propose a Predicting Center Module (PCM) that shares parameters with the original encoder with extra cross-attention to predict centers. Our method is of high pre-training efficiency compared to other alternatives and achieves great improvement over Point-MAE, particularly outperforming it by 5.50%, 6.03%, and 5.17% on three variants of ScanObjectNN. The code will be made publicly available.
Authors: Albert Gassol Puigjaner, Edoardo Mello Rella, Erik Sandstr\"om, Ajad Chhatkuli, Luc Van Gool
Abstract: Implicit surfaces via neural radiance fields (NeRF) have shown surprising accuracy in surface reconstruction. Despite their success in reconstructing richly textured surfaces, existing methods struggle with planar regions with weak textures, which account for the majority of indoor scenes. In this paper, we address indoor dense surface reconstruction by revisiting key aspects of NeRF in order to use the recently proposed Vector Field (VF) as the implicit representation. VF is defined by the unit vector directed to the nearest surface point. It therefore flips direction at the surface and equals to the explicit surface normals. Except for this flip, VF remains constant along planar surfaces and provides a strong inductive bias in representing planar surfaces. Concretely, we develop a novel density-VF relationship and a training scheme that allows us to learn VF via volume rendering By doing this, VF-NeRF can model large planar surfaces and sharp corners accurately. We show that, when depth cues are available, our method further improves and achieves state-of-the-art results in reconstructing indoor scenes and rendering novel views. We extensively evaluate VF-NeRF on indoor datasets and run ablations of its components.
Authors: Simone Ricci, Niccol\`o Biondi, Federico Pernici, Alberto Del Bimbo
Abstract: Visual retrieval systems face significant challenges when updating models with improved representations due to misalignment between the old and new representations. The costly and resource-intensive backfilling process involves recalculating feature vectors for images in the gallery set whenever a new model is introduced. To address this, prior research has explored backward-compatible training methods that enable direct comparisons between new and old representations without backfilling. Despite these advancements, achieving a balance between backward compatibility and the performance of independently trained models remains an open problem. In this paper, we address it by expanding the representation space with additional dimensions and learning an orthogonal transformation to achieve compatibility with old models and, at the same time, integrate new information. This transformation preserves the original feature space's geometry, ensuring that our model aligns with previous versions while also learning new data. Our Orthogonal Compatible Aligned (OCA) approach eliminates the need for re-indexing during model updates and ensures that features can be compared directly across different model updates without additional mapping functions. Experimental results on CIFAR-100 and ImageNet-1k demonstrate that our method not only maintains compatibility with previous models but also achieves state-of-the-art accuracy, outperforming several existing methods.
Authors: Rongxuan Wang, Xin Lu, Xiaoyang Liu, Xiaoyi Zou, Tongyi Cao, Ying Li
Abstract: Online vectorized High-Definition (HD) map construction is crucial for subsequent prediction and planning tasks in autonomous driving. Following MapTR paradigm, recent works have made noteworthy achievements. However, reference points are randomly initialized in mainstream methods, leading to unstable matching between predictions and ground truth. To address this issue, we introduce PriorMapNet to enhance online vectorized HD map construction with priors. We propose the PPS-Decoder, which provides reference points with position and structure priors. Fitted from the map elements in the dataset, prior reference points lower the learning difficulty and achieve stable matching. Furthermore, we propose the PF-Encoder to enhance the image-to-BEV transformation with BEV feature priors. Besides, we propose the DMD cross-attention, which decouples cross-attention along multi-scale and multi-sample respectively to achieve efficiency. Our proposed PriorMapNet achieves state-of-the-art performance in the online vectorized HD map construction task on nuScenes and Argoverse2 datasets. The code will be released publicly soon.
Authors: Lin Zhao, Xiao Chen, Eric Z. Chen, Yikang Liu, Terrence Chen, Shanhui Sun
Abstract: Medical image segmentation is crucial for clinical decision-making, but the scarcity of annotated data presents significant challenges. Few-shot segmentation (FSS) methods show promise but often require retraining on the target domain and struggle to generalize across different modalities. Similarly, adapting foundation models like the Segment Anything Model (SAM) for medical imaging has limitations, including the need for finetuning and domain-specific adaptation. To address these issues, we propose a novel method that adapts DINOv2 and Segment Anything Model 2 (SAM 2) for retrieval-augmented few-shot medical image segmentation. Our approach uses DINOv2's feature as query to retrieve similar samples from limited annotated data, which are then encoded as memories and stored in memory bank. With the memory attention mechanism of SAM 2, the model leverages these memories as conditions to generate accurate segmentation of the target image. We evaluated our framework on three medical image segmentation tasks, demonstrating superior performance and generalizability across various modalities without the need for any retraining or finetuning. Overall, this method offers a practical and effective solution for few-shot medical image segmentation and holds significant potential as a valuable annotation tool in clinical applications.
Authors: Guangyi Wang, Yuren Cai, Lijiang Li, Wei Peng, Songzhi Su
Abstract: Diffusion Probabilistic Models (DPMs) have shown remarkable potential in image generation, but their sampling efficiency is hindered by the need for numerous denoising steps. Most existing solutions accelerate the sampling process by proposing fast ODE solvers. However, the inevitable discretization errors of the ODE solvers are significantly magnified when the number of function evaluations (NFE) is fewer. In this work, we propose PFDiff, a novel training-free and orthogonal timestep-skipping strategy, which enables existing fast ODE solvers to operate with fewer NFE. Based on two key observations: a significant similarity in the model's outputs at time step size that is not excessively large during the denoising process of existing ODE solvers, and a high resemblance between the denoising process and SGD. PFDiff, by employing gradient replacement from past time steps and foresight updates inspired by Nesterov momentum, rapidly updates intermediate states, thereby reducing unnecessary NFE while correcting for discretization errors inherent in first-order ODE solvers. Experimental results demonstrate that PFDiff exhibits flexible applicability across various pre-trained DPMs, particularly excelling in conditional DPMs and surpassing previous state-of-the-art training-free methods. For instance, using DDIM as a baseline, we achieved 16.46 FID (4 NFE) compared to 138.81 FID with DDIM on ImageNet 64x64 with classifier guidance, and 13.06 FID (10 NFE) on Stable Diffusion with 7.5 guidance scale.
Authors: Andong Lu, Wanyu Wang, Chenglong Li, Jin Tang, Bin Luo
Abstract: Existing RGBT tracking methods often design various interaction models to perform cross-modal fusion of each layer, but can not execute the feature interactions among all layers, which plays a critical role in robust multimodal representation, due to large computational burden. To address this issue, this paper presents a novel All-layer multimodal Interaction Network, named AINet, which performs efficient and effective feature interactions of all modalities and layers in a progressive fusion Mamba, for robust RGBT tracking. Even though modality features in different layers are known to contain different cues, it is always challenging to build multimodal interactions in each layer due to struggling in balancing interaction capabilities and efficiency. Meanwhile, considering that the feature discrepancy between RGB and thermal modalities reflects their complementary information to some extent, we design a Difference-based Fusion Mamba (DFM) to achieve enhanced fusion of different modalities with linear complexity. When interacting with features from all layers, a huge number of token sequences (3840 tokens in this work) are involved and the computational burden is thus large. To handle this problem, we design an Order-dynamic Fusion Mamba (OFM) to execute efficient and effective feature interactions of all layers by dynamically adjusting the scan order of different layers in Mamba. Extensive experiments on four public RGBT tracking datasets show that AINet achieves leading performance against existing state-of-the-art methods.
Authors: Eman Ali, Sathira Silva, Muhammad Haris Khan
Abstract: Vision-language models (VLMs), e.g., CLIP, have shown remarkable potential in zero-shot image classification. However, adapting these models to new domains remains challenging, especially in unsupervised settings where labelled data is unavailable. Recent research has proposed pseudo-labelling approaches to adapt CLIP in an unsupervised manner using unlabelled target data. Nonetheless, these methods struggle due to noisy pseudo-labels resulting from the misalignment between CLIP's visual and textual representations. This study introduces DPA, an unsupervised domain adaptation method for VLMs. DPA introduces the concept of dual prototypes, acting as distinct classifiers, along with the convex combination of their outputs, thereby leading to accurate pseudo-label construction. Next, it ranks pseudo-labels to facilitate robust self-training, particularly during early training. Finally, it addresses visual-textual misalignment by aligning textual prototypes with image prototypes to further improve the adaptation performance. Experiments on 13 downstream vision tasks demonstrate that DPA significantly outperforms zero-shot CLIP and the state-of-the-art unsupervised adaptation baselines.
Authors: Xinyu Xiong, Zihuang Wu, Shuangyi Tan, Wenxue Li, Feilong Tang, Ying Chen, Siying Li, Jie Ma, Guanbin Li
Abstract: Image segmentation plays an important role in vision understanding. Recently, the emerging vision foundation models continuously achieved superior performance on various tasks. Following such success, in this paper, we prove that the Segment Anything Model 2 (SAM2) can be a strong encoder for U-shaped segmentation models. We propose a simple but effective framework, termed SAM2-UNet, for versatile image segmentation. Specifically, SAM2-UNet adopts the Hiera backbone of SAM2 as the encoder, while the decoder uses the classic U-shaped design. Additionally, adapters are inserted into the encoder to allow parameter-efficient fine-tuning. Preliminary experiments on various downstream tasks, such as camouflaged object detection, salient object detection, marine animal segmentation, mirror detection, and polyp segmentation, demonstrate that our SAM2-UNet can simply beat existing specialized state-of-the-art methods without bells and whistles. Project page: \url{https://github.com/WZH0120/SAM2-UNet}.
Authors: Le Xue, Manli Shu, Anas Awadalla, Jun Wang, An Yan, Senthil Purushwalkam, Honglu Zhou, Viraj Prabhu, Yutong Dai, Michael S Ryoo, Shrikant Kendre, Jieyu Zhang, Can Qin, Shu Zhang, Chia-Chih Chen, Ning Yu, Juntao Tan, Tulika Manoj Awalgaonkar, Shelby Heinecke, Huan Wang, Yejin Choi, Ludwig Schmidt, Zeyuan Chen, Silvio Savarese, Juan Carlos Niebles, Caiming Xiong, Ran Xu
Abstract: This report introduces xGen-MM (also known as BLIP-3), a framework for developing Large Multimodal Models (LMMs). The framework comprises meticulously curated datasets, a training recipe, model architectures, and a resulting suite of LMMs. xGen-MM, short for xGen-MultiModal, expands the Salesforce xGen initiative on foundation AI models. Our models undergo rigorous evaluation across a range of tasks, including both single and multi-image benchmarks. Our pre-trained base model exhibits strong in-context learning capabilities and the instruction-tuned model demonstrates competitive performance among open-source LMMs with similar model sizes. In addition, we introduce a safety-tuned model with DPO, aiming to mitigate harmful behaviors such as hallucinations and improve safety. We open-source our models, curated large-scale datasets, and our fine-tuning codebase to facilitate further advancements in LMM research. Associated resources will be available on our project page above.
Authors: Junxian He, Shrinivas Pundlik, Gang Luo
Abstract: Objective: Micro-navigation poses challenges for blind and visually impaired individuals. They often need to ask for sighted assistance. We explored the feasibility of utilizing ChatGPT as a virtual assistant to provide navigation directions. Methods: We created a test set of outdoor and indoor micro-navigation scenarios consisting of 113 scene images and their human-generated text descriptions. A total of 412 way-finding queries and their expected responses were compiled based on the scenarios. Not all queries are answerable based on the information available in the scene image. "I do not know"response was expected for unanswerable queries, which served as negative cases. High level orientation responses were expected, and step-by-step guidance was not required. ChatGPT 4o was evaluated based on sensitivity (SEN) and specificity (SPE) under different conditions. Results: The default ChatGPT 4o, with scene images as inputs, resulted in SEN and SPE values of 64.8% and 75.9%, respectively. Instruction on how to respond to unanswerable questions did not improve SEN substantially but SPE increased by around 14 percentage points. SEN and SPE both improved substantially, by about 17 and 16 percentage points on average respectively, when human written descriptions of the scenes were provided as input instead of images. Providing further prompt instructions to the assistants when the input was text description did not substantially change the SEN and SPE values. Conclusion: Current native ChatGPT 4o is still unable to provide correct micro-navigation guidance in some cases, probably because its scene understanding is not optimized for navigation purposes. If multi-modal chatbots could interpret scenes with a level of clarity comparable to humans, and also guided by appropriate prompts, they may have the potential to provide assistance to visually impaired for micro-navigation.
Authors: Tomasz Prytu{\l}a
Abstract: We give an overview of combinatorial methods to represent 3D data, such as graphs and meshes, from the viewpoint of their amenability to analysis using machine learning algorithms. We highlight pros and cons of various representations and we discuss some methods of generating/switching between the representations. We finally present two concrete applications in life science and industry. Despite its theoretical nature, our discussion is in general motivated by, and biased towards real-world challenges.
Authors: Alexander Varlamov, Daria Diatlova, Egor Spirin
Abstract: Improvements in diffusion models have boosted the quality of image generation, which has led researchers, companies, and creators to focus on improving watermarking algorithms. This provision would make it possible to clearly identify the creators of generative art. The main challenges that modern watermarking algorithms face have to do with their ability to withstand attacks and encrypt many unique messages, such as user IDs. In this paper, we present METR: Message Enhanced Tree-Ring, which is an approach that aims to address these challenges. METR is built on the Tree-Ring watermarking algorithm, a technique that makes it possible to encode multiple distinct messages without compromising attack resilience or image quality. This ensures the suitability of this watermarking algorithm for any Diffusion Model. In order to surpass the limitations on the quantity of encoded messages, we propose METR++, an enhanced version of METR. This approach, while limited to the Latent Diffusion Model architecture, is designed to inject a virtually unlimited number of unique messages. We demonstrate its robustness to attacks and ability to encrypt many unique messages while preserving image quality, which makes METR and METR++ hold great potential for practical applications in real-world settings. Our code is available at https://github.com/deepvk/metr
Authors: Ce Chen, Shaoli Huang, Xuelin Chen, Guangyi Chen, Xiaoguang Han, Kun Zhang, Mingming Gong
Abstract: Text-to-4D generation has recently been demonstrated viable by integrating a 2D image diffusion model with a video diffusion model. However, existing models tend to produce results with inconsistent motions and geometric structures over time. To this end, we present a novel framework, coined CT4D, which directly operates on animatable meshes for generating consistent 4D content from arbitrary user-supplied prompts. The primary challenges of our mesh-based framework involve stably generating a mesh with details that align with the text prompt while directly driving it and maintaining surface continuity. Our CT4D framework incorporates a unique Generate-Refine-Animate (GRA) algorithm to enhance the creation of text-aligned meshes. To improve surface continuity, we divide a mesh into several smaller regions and implement a uniform driving function within each area. Additionally, we constrain the animating stage with a rigidity regulation to ensure cross-region continuity. Our experimental results, both qualitative and quantitative, demonstrate that our CT4D framework surpasses existing text-to-4D techniques in maintaining interframe consistency and preserving global geometry. Furthermore, we showcase that this enhanced representation inherently possesses the capability for combinational 4D generation and texture editing.
Authors: Abdur R. Fayjie, Jutika Borah, Florencia Carbone, Jan Tack, Patrick Vandewalle
Abstract: Deep learning has shown tremendous progress in a wide range of digital pathology and medical image classification tasks. Its integration into safe clinical decision-making support requires robust and reliable models. However, real-world data comes with diversities that often lie outside the intended source distribution. Moreover, when test samples are dramatically different, clinical decision-making is greatly affected. Quantifying predictive uncertainty in models is crucial for well-calibrated predictions and determining when (or not) to trust a model. Unfortunately, many works have overlooked the importance of predictive uncertainty estimation. This paper evaluates whether predictive uncertainty estimation adds robustness to deep learning-based diagnostic decision-making systems. We investigate the effect of various carcinoma distribution shift scenarios on predictive performance and calibration. We first systematically investigate three popular methods for improving predictive uncertainty: Monte Carlo dropout, deep ensemble, and few-shot learning on lung adenocarcinoma classification as a primary disease in whole slide images. Secondly, we compare the effectiveness of the methods in terms of performance and calibration under clinically relevant distribution shifts such as in-distribution shifts comprising primary disease sub-types and other characterization analysis data; out-of-distribution shifts comprising well-differentiated cases, different organ origin, and imaging modality shifts. While studies on uncertainty estimation exist, to our best knowledge, no rigorous large-scale benchmark compares predictive uncertainty estimation including these dataset shifts for lung carcinoma classification.
Authors: Yusen Wu, Hao Chen, Alex Pissinou Makki, Phuong Nguyen, Yelena Yesha
Abstract: Distributional drift detection is important in medical applications as it helps ensure the accuracy and reliability of models by identifying changes in the underlying data distribution that could affect diagnostic or treatment decisions. However, current methods have limitations in detecting drift; for example, the inclusion of abnormal datasets can lead to unfair comparisons. This paper presents an accurate and sensitive approach to detect distributional drift in CT-scan medical images by leveraging data-sketching and fine-tuning techniques. We developed a robust baseline library model for real-time anomaly detection, allowing for efficient comparison of incoming images and identification of anomalies. Additionally, we fine-tuned a vision transformer pre-trained model to extract relevant features using breast cancer images as an example, significantly enhancing model accuracy to 99.11\%. Combining with data-sketches and fine-tuning, our feature extraction evaluation demonstrated that cosine similarity scores between similar datasets provide greater improvements, from around 50\% increased to 100\%. Finally, the sensitivity evaluation shows that our solutions are highly sensitive to even 1\% salt-and-pepper and speckle noise, and it is not sensitive to lighting noise (e.g., lighting conditions have no impact on data drift). The proposed methods offer a scalable and reliable solution for maintaining the accuracy of diagnostic models in dynamic clinical environments.
Authors: Xiaochuang Han, Marjan Ghazvininejad, Pang Wei Koh, Yulia Tsvetkov
Abstract: Recent work in image and video generation has been adopting the autoregressive LLM architecture due to its generality and potentially easy integration into multi-modal systems. The crux of applying autoregressive training in language generation to visual generation is discretization -- representing continuous data like images and videos as discrete tokens. Common methods of discretizing images and videos include modeling raw pixel values, which are prohibitively lengthy, or vector quantization, which requires convoluted pre-hoc training. In this work, we propose to directly model images and videos as compressed files saved on computers via canonical codecs (e.g., JPEG, AVC/H.264). Using the default Llama architecture without any vision-specific modifications, we pretrain JPEG-LM from scratch to generate images (and AVC-LM to generate videos as a proof of concept), by directly outputting compressed file bytes in JPEG and AVC formats. Evaluation of image generation shows that this simple and straightforward approach is more effective than pixel-based modeling and sophisticated vector quantization baselines (on which our method yields a 31% reduction in FID). Our analysis shows that JPEG-LM has an especial advantage over vector quantization models in generating long-tail visual elements. Overall, we show that using canonical codec representations can help lower the barriers between language generation and visual generation, facilitating future research on multi-modal language/image/video LLMs.
Authors: Mohammad Hossein Najafi, Mohammad Morsali, Mohammadmahdi Vahediahmar, Saeed Bagheri Shouraki
Abstract: Recent advancements in deep learning, particularly in medical imaging, have significantly propelled the progress of healthcare systems. However, examining the robustness of medical images against adversarial attacks is crucial due to their real-world applications and profound impact on individuals' health. These attacks can result in misclassifications in disease diagnosis, potentially leading to severe consequences. Numerous studies have explored both the implementation of adversarial attacks on medical images and the development of defense mechanisms against these threats, highlighting the vulnerabilities of deep neural networks to such adversarial activities. In this study, we investigate adversarial attacks on images associated with Alzheimer's disease and propose a defensive method to counteract these attacks. Specifically, we examine adversarial attacks that employ frequency domain transformations on Alzheimer's disease images, along with other well-known adversarial attacks. Our approach utilizes a convolutional neural network (CNN)-based autoencoder architecture in conjunction with the two-dimensional Fourier transform of images for detection purposes. The simulation results demonstrate that our detection and defense mechanism effectively mitigates several adversarial attacks, thereby enhancing the robustness of deep neural networks against such vulnerabilities.
Authors: S\'andor Gazdag, Tom M\"oller, Tam\'as Filep, Anita Keszler, Andr\'as L. Majdik
Abstract: The usage of commercial Micro Aerial Vehicles (MAVs) has increased drastically during the last decade. While the added value of MAVs to society is apparent, their growing use is also coming with increasing risks like violating public airspace at airports or committing privacy violations. To mitigate these issues it is becoming critical to develop solutions that incorporate the detection and tracking of MAVs with autonomous systems. This work presents a method for the detection and tracking of MAVs using a novel, low-cost rosette scanning LiDAR on a pan-tilt turret. Once the static background is captured, a particle filter is utilized to detect a possible target and track its position with a physical, programmable pan-tilt system. The tracking makes it possible to keep the MAV in the center, maximizing the density of 3D points measured on the target by the LiDAR sensor. The developed algorithm was evaluated within the indoor MIcro aerial vehicle and MOtion capture (MIMO) arena and has state-of-the-art tracking accuracy, stability, and fast re-detection time in case of tracking loss. Based on the outdoor tests, it was possible to significantly increase the detection distance and number of returned points compared to other similar methods using LiDAR.
Authors: Xue Wang, Tian Zhou, Jianqing Zhu, Jialin Liu, Kun Yuan, Tao Yao, Wotao Yin, Rong Jin, HanQin Cai
Abstract: Attention based models have achieved many remarkable breakthroughs in numerous applications. However, the quadratic complexity of Attention makes the vanilla Attention based models hard to apply to long sequence tasks. Various improved Attention structures are proposed to reduce the computation cost by inducing low rankness and approximating the whole sequence by sub-sequences. The most challenging part of those approaches is maintaining the proper balance between information preservation and computation reduction: the longer sub-sequences used, the better information is preserved, but at the price of introducing more noise and computational costs. In this paper, we propose a smoothed skeleton sketching based Attention structure, coined S$^3$Attention, which significantly improves upon the previous attempts to negotiate this trade-off. S$^3$Attention has two mechanisms to effectively minimize the impact of noise while keeping the linear complexity to the sequence length: a smoothing block to mix information over long sequences and a matrix sketching method that simultaneously selects columns and rows from the input matrix. We verify the effectiveness of S$^3$Attention both theoretically and empirically. Extensive studies over Long Range Arena (LRA) datasets and six time-series forecasting show that S$^3$Attention significantly outperforms both vanilla Attention and other state-of-the-art variants of Attention structures.
Authors: Daniel Omeiza, Pratik Somaiya, Jo-Ann Pattinson, Carolyn Ten-Holter, Jack Stilgoe, Marina Jirotka, Lars Kunze
Abstract: As artificial intelligence (AI) technology advances, ensuring the robustness and safety of AI-driven systems has become paramount. However, varying perceptions of robustness among AI developers create misaligned evaluation metrics, complicating the assessment and certification of safety-critical and complex AI systems such as autonomous driving (AD) agents. To address this challenge, we introduce Simulation-Based Robustness Assessment Framework (S-RAF) for autonomous driving. S-RAF leverages the CARLA Driving simulator to rigorously assess AD agents across diverse conditions, including faulty sensors, environmental changes, and complex traffic situations. By quantifying robustness and its relationship with other safety-critical factors, such as carbon emissions, S-RAF aids developers and stakeholders in building safe and responsible driving agents, and streamlining safety certification processes. Furthermore, S-RAF offers significant advantages, such as reduced testing costs, and the ability to explore edge cases that may be unsafe to test in the real world. The code for this framework is available here: https://github.com/cognitive-robots/rai-leaderboard
Authors: Kyungryun Lee, Won-Ki Jeong
Abstract: Analysis and visualization of 3D microscopy images pose challenges due to anisotropic axial resolution, demanding volumetric super-resolution along the axial direction. While training a learning-based 3D super-resolution model seems to be a straightforward solution, it requires ground truth isotropic volumes and suffers from the curse of dimensionality. Therefore, existing methods utilize 2D neural networks to reconstruct each axial slice, eventually piecing together the entire volume. However, reconstructing each slice in the pixel domain fails to give consistent reconstruction in all directions leading to misalignment artifacts. In this work, we present a reconstruction framework based on implicit neural representation (INR), which allows 3D coherency even when optimized by independent axial slices in a batch-wise manner. Our method optimizes a continuous volumetric representation from low-resolution axial slices, using a 2D diffusion prior trained on high-resolution lateral slices without requiring isotropic volumes. Through experiments on real and synthetic anisotropic microscopy images, we demonstrate that our method surpasses other state-of-the-art reconstruction methods. The source code is available on GitHub: https://github.com/hvcl/INR-diffusion.
Authors: Jian Li, Weiheng Lu
Abstract: Multimodal Large Language Models (MLLMs) are gaining increasing popularity in both academia and industry due to their remarkable performance in various applications such as visual question answering, visual perception, understanding, and reasoning. Over the past few years, significant efforts have been made to examine MLLMs from multiple perspectives. This paper presents a comprehensive review of \textbf{180 benchmarks} and evaluation for MLLMs, focusing on (1)perception and understanding, (2)cognition and reasoning, (3)specific domains, (4)key capabilities, and (5)other modalities. Finally, we discuss the limitations of the current evaluation methods for MLLMs and explore promising future directions. Our key argument is that evaluation should be regarded as a crucial discipline to better support the development of MLLMs. For more details, please visit our GitHub repository: https://github.com/swordlidev/Evaluation-Multimodal-LLMs-Survey.
URLs: https://github.com/swordlidev/Evaluation-Multimodal-LLMs-Survey.
Authors: Florentin Bieder, Paul Friedrich, H\'el\`ene Corbaz, Alicia Durrer, Julia Wolleb, Philippe C. Cattin
Abstract: The human brain undergoes rapid development during the third trimester of pregnancy. In this work, we model the neonatal development of the infant brain in this age range. As a basis, we use MR images of preterm- and term-birth neonates from the developing human connectome project (dHCP). We propose a neural network, specifically an implicit neural representation (INR), to predict 2D- and 3D images of varying time points. In order to model a subject-specific development process, it is necessary to disentangle the age from the subjects' identity in the latent space of the INR. We propose two methods, Subject Specific Latent Vectors (SSL) and Stochastic Global Latent Augmentation (SGLA), enabling this disentanglement. We perform an analysis of the results and compare our proposed model to an age-conditioned denoising diffusion model as a baseline. We also show that our method can be applied in a memory-efficient way, which is especially important for 3D data.
Authors: Qichen Zheng, Yi Yu, Siyuan Yang, Jun Liu, Kwok-Yan Lam, Alex Kot
Abstract: Skeleton Action Recognition (SAR) has attracted significant interest for its efficient representation of the human skeletal structure. Despite its advancements, recent studies have raised security concerns in SAR models, particularly their vulnerability to adversarial attacks. However, such strategies are limited to digital scenarios and ineffective in physical attacks, limiting their real-world applicability. To investigate the vulnerabilities of SAR in the physical world, we introduce the Physical Skeleton Backdoor Attacks (PSBA), the first exploration of physical backdoor attacks against SAR. Considering the practicalities of physical execution, we introduce a novel trigger implantation method that integrates infrequent and imperceivable actions as triggers into the original skeleton data. By incorporating a minimal amount of this manipulated data into the training set, PSBA enables the system misclassify any skeleton sequences into the target class when the trigger action is present. We examine the resilience of PSBA in both poisoned and clean-label scenarios, demonstrating its efficacy across a range of datasets, poisoning ratios, and model architectures. Additionally, we introduce a trigger-enhancing strategy to strengthen attack performance in the clean label setting. The robustness of PSBA is tested against three distinct backdoor defenses, and the stealthiness of PSBA is evaluated using two quantitative metrics. Furthermore, by employing a Kinect V2 camera, we compile a dataset of human actions from the real world to mimic physical attack situations, with our findings confirming the effectiveness of our proposed attacks. Our project website can be found at https://qichenzheng.github.io/psba-website.
Authors: Yuqi Ye, Wei Gao
Abstract: The key to effective point cloud compression is to obtain a robust context model consistent with complex 3D data structures. Recently, the advancement of large language models (LLMs) has highlighted their capabilities not only as powerful generators for in-context learning and generation but also as effective compressors. These dual attributes of LLMs make them particularly well-suited to meet the demands of data compression. Therefore, this paper explores the potential of using LLM for compression tasks, focusing on lossless point cloud geometry compression (PCGC) experiments. However, applying LLM directly to PCGC tasks presents some significant challenges, i.e., LLM does not understand the structure of the point cloud well, and it is a difficult task to fill the gap between text and point cloud through text description, especially for large complicated and small shapeless point clouds. To address these problems, we introduce a novel architecture, namely the Large Language Model-based Point Cloud Geometry Compression (LLM-PCGC) method, using LLM to compress point cloud geometry information without any text description or aligning operation. By utilizing different adaptation techniques for cross-modality representation alignment and semantic consistency, including clustering, K-tree, token mapping invariance, and Low Rank Adaptation (LoRA), the proposed method can translate LLM to a compressor/generator for point cloud. To the best of our knowledge, this is the first structure to employ LLM as a compressor for point cloud data. Experiments demonstrate that the LLM-PCGC outperforms the other existing methods significantly, by achieving -40.213% bit rate reduction compared to the reference software of MPEG Geometry-based Point Cloud Compression (G-PCC) standard, and by achieving -2.267% bit rate reduction compared to the state-of-the-art learning-based method.
Authors: Xiaoyu Wang, Martin Benning, Audrey Repetti
Abstract: Unfolded proximal neural networks (PNNs) form a family of methods that combines deep learning and proximal optimization approaches. They consist in designing a neural network for a specific task by unrolling a proximal algorithm for a fixed number of iterations, where linearities can be learned from prior training procedure. PNNs have shown to be more robust than traditional deep learning approaches while reaching at least as good performances, in particular in computational imaging. However, training PNNs still depends on the efficiency of available training algorithms. In this work, we propose a lifted training formulation based on Bregman distances for unfolded PNNs. Leveraging the deterministic mini-batch block-coordinate forward-backward method, we design a bespoke computational strategy beyond traditional back-propagation methods for solving the resulting learning problem efficiently. We assess the behaviour of the proposed training approach for PNNs through numerical simulations on image denoising, considering a denoising PNN whose structure is based on dual proximal-gradient iterations.
Authors: Ashesh Ashesh, Joran Deschamps, Florian Jug
Abstract: Microscopy is routinely used to image biological structures of interest. Due to imaging constraints, acquired images are typically low-SNR and contain noise. Over the last few years, regression-based tasks like unsupervised denoising and splitting have found utility in working with such noisy micrographs. For evaluation, Structural Similarity (SSIM) is one of the most popular measures used in the field. For such tasks, the best evaluation would be when both low-SNR noisy images and corresponding high-SNR clean images are obtained directly from a microscope. However, due to the following three peculiar properties of the microscopy data, we observe that SSIM is not well suited to this data regime: (a) high-SNR micrographs have higher intensity pixels as compared to low SNR micrographs, (b) high-SNR micrographs have higher intensity pixels than found in natural images, images for which SSIM was developed, and (c) a digitally configurable offset is added by the detector present inside the microscope. We show that SSIM components behave unexpectedly when the prediction generated from low-SNR input is compared with the corresponding high-SNR data. We explain this behavior by introducing the phenomenon of saturation, where the value of SSIM components becomes less sensitive to (dis)similarity between the images. We introduce microSSIM, a variant of SSIM, which overcomes the above-discussed issues. We justify the soundness and utility of microSSIM using theoretical and empirical arguments and show the utility of microSSIM on two tasks: unsupervised denoising and joint image splitting with unsupervised denoising. Since our formulation can be applied to a broad family of SSIM-based measures, we also introduce MicroMS3IM, a microscopy-specific variation of MS-SSIM. The source code and python package is available at https://github.com/juglab/MicroSSIM.
Authors: Miriam Cobo, Amaia P\'erez del Barrio, Pablo Men\'endez Fern\'andez-Miranda, Pablo Sanz Bell\'on, Lara Lloret Iglesias, Wilson Silva
Abstract: Prognosis after intracranial hemorrhage (ICH) is influenced by a complex interplay between imaging and tabular data. Rapid and reliable prognosis are crucial for effective patient stratification and informed treatment decision-making. In this study, we aim to enhance image-based prognosis by learning a robust feature representation shared between prognosis and the clinical and demographic variables most highly correlated with it. Our approach mimics clinical decision-making by reinforcing the model to learn valuable prognostic data embedded in the image. We propose a 3D multi-task image model to predict prognosis, Glasgow Coma Scale and age, improving accuracy and interpretability. Our method outperforms current state-of-the-art baseline image models, and demonstrates superior performance in ICH prognosis compared to four board-certified neuroradiologists using only CT scans as input. We further validate our model with interpretability saliency maps. Code is available at https://github.com/MiriamCobo/MultitaskLearning_ICH_Prognosis.git.
URLs: https://github.com/MiriamCobo/MultitaskLearning_ICH_Prognosis.git.
Authors: Boa Jang, Youngbin Ahn, Eun Kyung Choe, Chang Ki Yoon, Hyuk Jin Choi, Young-Gon Kim
Abstract: Artificial intelligence applied to retinal images offers significant potential for recognizing signs and symptoms of retinal conditions and expediting the diagnosis of eye diseases and systemic disorders. However, developing generalized artificial intelligence models for medical data often requires a large number of labeled images representing various disease signs, and most models are typically task-specific, focusing on major retinal diseases. In this study, we developed a Fundus-Specific Pretrained Model (Image+Fundus), a supervised artificial intelligence model trained to detect abnormalities in fundus images. A total of 57,803 images were used to develop this pretrained model, which achieved superior performance across various downstream tasks, indicating that our proposed model outperforms other general methods. Our Image+Fundus model offers a generalized approach to improve model performance while reducing the number of labeled datasets required. Additionally, it provides more disease-specific insights into fundus images, with visualizations generated by our model. These disease-specific foundation models are invaluable in enhancing the performance and efficiency of deep learning models in the field of fundus imaging.
Authors: Louise Guillon, Soheib Biga, Axel Puyo, Gr\'egoire Pasquier, Valentin Foucher, Yendoub\'e E. Kantchire, St\'ephane E. Sossou, Ameyo M. Dorkenoo, Laurent Bonnardot, Marc Thellier, Laurence Lachaud, Renaud Piarroux
Abstract: Malaria remains a significant global health challenge, necessitating rapid and accurate diagnostic methods. While computer-aided diagnosis (CAD) tools utilizing deep learning have shown promise, their generalization to diverse clinical settings remains poorly assessed. This study evaluates the generalization capabilities of a CAD model for malaria diagnosis from thin blood smear images across four sites. We explore strategies to enhance generalization, including fine-tuning and incremental learning. Our results demonstrate that incorporating site-specific data significantly improves model performance, paving the way for broader clinical application.
Authors: Zhi-Bo Liu, Xiaobo Pang, Jizhao Wang, Shuai Liu, Chen Li
Abstract: In pathological research, education, and clinical practice, the decision-making process based on pathological images is critically important. This significance extends to digital pathology image analysis: its adequacy is demonstrated by the extensive information contained within tissue structures, which is essential for accurate cancer classification and grading. Additionally, its necessity is highlighted by the inherent requirement for interpretability in the conclusions generated by algorithms. For humans, determining tumor type and grade typically involves multi-scale analysis, which presents a significant challenge for AI algorithms. Traditional patch-based methods are inadequate for modeling such complex structures, as they fail to capture the intricate, multi-scale information inherent in whole slide images. Consequently, there is a pressing need for advanced AI techniques capable of efficiently and accurately replicating this complex analytical process. To address this issue, we introduce HistoGym, an open-source reinforcement learning environment for histopathological image analysis. Following OpenAI Gym APIs, HistoGym aims to foster whole slide image diagnosis by mimicking the real-life processes of doctors. Leveraging the pyramid feature of WSIs and the OpenSlide API, HistoGym provides a unified framework for various clinical tasks, including tumor detection and classification. We detail the observation, action, and reward specifications tailored for the histopathological image analysis domain and provide an open-source Python-based interface for both clinicians and researchers. To accommodate different clinical demands, we offer various scenarios for different organs and cancers, including both WSI-based and selected region-based scenarios, showcasing several noteworthy results.
Authors: Junjie Liu, Junlong Liu, Rongxin Jiang, Yaowu Chen, Chen Shen, Jieping Ye
Abstract: Multi-view Clustering (MVC) has achieved significant progress, with many efforts dedicated to learn knowledge from multiple views. However, most existing methods are either not applicable or require additional steps for incomplete MVC. Such a limitation results in poor-quality clustering performance and poor missing view adaptation. Besides, noise or outliers might significantly degrade the overall clustering performance, which are not handled well by most existing methods. In this paper, we propose a novel unified framework for incomplete and complete MVC named self-learning symmetric multi-view probabilistic clustering (SLS-MPC). SLS-MPC proposes a novel symmetric multi-view probability estimation and equivalently transforms multi-view pairwise posterior matching probability into composition of each view's individual distribution, which tolerates data missing and might extend to any number of views. Then, SLS-MPC proposes a novel self-learning probability function without any prior knowledge and hyper-parameters to learn each view's individual distribution. Next, graph-context-aware refinement with path propagation and co-neighbor propagation is used to refine pairwise probability, which alleviates the impact of noise and outliers. Finally, SLS-MPC proposes a probabilistic clustering algorithm to adjust clustering assignments by maximizing the joint probability iteratively without category information. Extensive experiments on multiple benchmarks show that SLS-MPC outperforms previous state-of-the-art methods.
Authors: Yi Shi, Jingbo Wang, Xuekun Jiang, Bingkun Lin, Bo Dai, Xue Bin Peng
Abstract: Real-time character control is an essential component for interactive experiences, with a broad range of applications, including physics simulations, video games, and virtual reality. The success of diffusion models for image synthesis has led to the use of these models for motion synthesis. However, the majority of these motion diffusion models are primarily designed for offline applications, where space-time models are used to synthesize an entire sequence of frames simultaneously with a pre-specified length. To enable real-time motion synthesis with diffusion model that allows time-varying controls, we propose A-MDM (Auto-regressive Motion Diffusion Model). Our conditional diffusion model takes an initial pose as input, and auto-regressively generates successive motion frames conditioned on the previous frame. Despite its streamlined network architecture, which uses simple MLPs, our framework is capable of generating diverse, long-horizon, and high-fidelity motion sequences. Furthermore, we introduce a suite of techniques for incorporating interactive controls into A-MDM, such as task-oriented sampling, in-painting, and hierarchical reinforcement learning. These techniques enable a pre-trained A-MDM to be efficiently adapted for a variety of new downstream tasks. We conduct a comprehensive suite of experiments to demonstrate the effectiveness of A-MDM, and compare its performance against state-of-the-art auto-regressive methods.
Authors: Shengkun Tang, Yaqing Wang, Caiwen Ding, Yi Liang, Yao Li, Dongkuan Xu
Abstract: Diffusion models achieve great success in generating diverse and high-fidelity images, yet their widespread application, especially in real-time scenarios, is hampered by their inherently slow generation speed. The slow generation stems from the necessity of multi-step network inference. While some certain predictions benefit from the full computation of the model in each sampling iteration, not every iteration requires the same amount of computation, potentially leading to inefficient computation. Unlike typical adaptive computation challenges that deal with single-step generation problems, diffusion processes with a multi-step generation need to dynamically adjust their computational resource allocation based on the ongoing assessment of each step's importance to the final image output, presenting a unique set of challenges. In this work, we propose AdaDiff, an adaptive framework that dynamically allocates computation resources in each sampling step to improve the generation efficiency of diffusion models. To assess the effects of changes in computational effort on image quality, we present a timestep-aware uncertainty estimation module (UEM). Integrated at each intermediate layer, the UEM evaluates the predictive uncertainty. This uncertainty measurement serves as an indicator for determining whether to terminate the inference process. Additionally, we introduce an uncertainty-aware layer-wise loss aimed at bridging the performance gap between full models and their adaptive counterparts.
Authors: Chong Zhong, Yang Li, Danjuan Yang, Meiyan Li, Xingyao Zhou, Bo Fu, Catherine C. Liu, A. H. Welsh
Abstract: The ultra-widefield (UWF) fundus image is an attractive 3D biomarker in AI-aided myopia screening because it provides much richer myopia-related information. Though axial length (AL) has been acknowledged to be highly related to the two key targets of myopia screening, Spherical Equivalence (SE) measurement and high myopia diagnosis, its prediction based on the UWF fundus image is rarely considered. To save the high expense and time costs of measuring SE and AL, we propose the Copula-enhanced Convolutional Neural Network (CeCNN), a one-stop UWF-based ophthalmic AI framework to jointly predict SE, AL, and myopia status. The CeCNN formulates a multiresponse regression that relates multiple dependent discrete-continuous responses and the image covariate, where the nonlinearity of the association is modeled by a backbone CNN. To thoroughly describe the dependence structure among the responses, we model and incorporate the conditional dependence among responses in a CNN through a new copula-likelihood loss. We provide statistical interpretations of the conditional dependence among responses, and reveal that such dependence is beyond the dependence explained by the image covariate. We heuristically justify that the proposed loss can enhance the estimation efficiency of the CNN weights. We apply the CeCNN to the UWF dataset collected by us and demonstrate that the CeCNN sharply enhances the predictive capability of various backbone CNNs. Our study evidences the ophthalmology view that besides SE, AL is also an important measure to myopia.
Authors: Wenquan Lu, Yufei Xu, Jing Zhang, Chaoyue Wang, Dacheng Tao
Abstract: Diffusion models have achieved remarkable success in generating realistic images but suffer from generating accurate human hands, such as incorrect finger counts or irregular shapes. This difficulty arises from the complex task of learning the physical structure and pose of hands from training images, which involves extensive deformations and occlusions. For correct hand generation, our paper introduces a lightweight post-processing solution called $\textbf{HandRefiner}$. HandRefiner employs a conditional inpainting approach to rectify malformed hands while leaving other parts of the image untouched. We leverage the hand mesh reconstruction model that consistently adheres to the correct number of fingers and hand shape, while also being capable of fitting the desired hand pose in the generated image. Given a generated failed image due to malformed hands, we utilize ControlNet modules to re-inject such correct hand information. Additionally, we uncover a phase transition phenomenon within ControlNet as we vary the control strength. It enables us to take advantage of more readily available synthetic data without suffering from the domain gap between realistic and synthetic hands. Experiments demonstrate that HandRefiner can significantly improve the generation quality quantitatively and qualitatively. The code is available at https://github.com/wenquanlu/HandRefiner .
Authors: Xin Lin, Jingtong Yue, Kelvin C. K. Chan, Lu Qi, Chao Ren, Jinshan Pan, Ming-Hsuan Yang
Abstract: Multi-task image restoration has gained significant interest due to its inherent versatility and efficiency compared to its single-task counterpart. However, performance decline is observed with an increase in the number of tasks, primarily attributed to the restoration model's challenge in handling different tasks with distinct natures at the same time. Thus, a perspective emerged aiming to explore the degradation-insensitive semantic commonalities among different degradation tasks. In this paper, we observe that the features of DINOv2 can effectively model semantic information and are independent of degradation factors. Motivated by this observation, we propose \mbox{\textbf{DINO-IR}}, a multi-task image restoration approach leveraging robust features extracted from DINOv2 to solve multi-task image restoration simultaneously. We first propose a pixel-semantic fusion (PSF) module to dynamically fuse DINOV2's shallow features containing pixel-level information and deep features containing degradation-independent semantic information. To guide the restoration model with the features of DINOv2, we develop a DINO-Restore adaption and fusion module to adjust the channel of fused features from PSF and then integrate them with the features from the restoration model. By formulating these modules into a unified deep model, we propose a DINO perception contrastive loss to constrain the model training. Extensive experimental results demonstrate that our DINO-IR performs favorably against existing multi-task image restoration approaches in various tasks by a large margin. The source codes and trained models will be made available.
Authors: Xiaoyun Xu, Shujian Yu, Zhuoran Liu, Stjepan Picek
Abstract: Vision Transformers (ViTs) achieve excellent performance in various tasks, but they are also vulnerable to adversarial attacks. Building robust ViTs is highly dependent on dedicated Adversarial Training (AT) strategies. However, current ViTs' adversarial training only employs well-established training approaches from convolutional neural network (CNN) training, where pre-training provides the basis for AT fine-tuning with the additional help of tailored data augmentations. In this paper, we take a closer look at the adversarial robustness of ViTs by providing a novel theoretical Mutual Information (MI) analysis in its autoencoder-based self-supervised pre-training. Specifically, we show that MI between the adversarial example and its latent representation in ViT-based autoencoders should be constrained by utilizing the MI bounds. Based on this finding, we propose a masked autoencoder-based pre-training method, MIMIR, that employs an MI penalty to facilitate the adversarial training of ViTs. Extensive experiments show that MIMIR outperforms state-of-the-art adversarially trained ViTs on benchmark datasets with higher natural and robust accuracy, indicating that ViTs can substantially benefit from exploiting MI. In addition, we consider two adaptive attacks by assuming that the adversary is aware of the MIMIR design, which further verifies the provided robustness.
Authors: Yuhao Jia, Wenhan Tan
Abstract: Diffusion-driven text-to-image (T2I) generation has achieved remarkable advancements. To further improve T2I models' capability in numerical and spatial reasoning, the layout is employed as an intermedium to bridge large language models and layout-based diffusion models. However, these methods still struggle with generating images from textural prompts with multiple objects and complicated spatial relationships. To tackle this challenge, we introduce a divide-and-conquer approach which decouples the T2I generation task into simple subtasks. Our approach divides the layout prediction stage into numerical & spatial reasoning and bounding box prediction. Then, the layout-to-image generation stage is conducted in an iterative manner to reconstruct objects from easy ones to difficult ones. We conduct experiments on the HRS and NSR-1K benchmarks and our approach outperforms previous state-of-the-art models with notable margins. In addition, visual results demonstrate that our approach significantly improves the controllability and consistency in generating multiple objects from complex textural prompts.
Authors: Yiming Ma, Victor Sanchez, Tanaya Guha
Abstract: We propose CLIP-EBC, the first fully CLIP-based model for accurate crowd density estimation. While the CLIP model has demonstrated remarkable success in addressing recognition tasks such as zero-shot image classification, its potential for counting has been largely unexplored due to the inherent challenges in transforming a regression problem, such as counting, into a recognition task. In this work, we investigate and enhance CLIP's ability to count, focusing specifically on the task of estimating crowd sizes from images. Existing classification-based crowd-counting frameworks have significant limitations, including the quantization of count values into bordering real-valued bins and the sole focus on classification errors. These practices result in label ambiguity near the shared borders and inaccurate prediction of count values. Hence, directly applying CLIP within these frameworks may yield suboptimal performance. To address these challenges, we first propose the Enhanced Blockwise Classification (EBC) framework. Unlike previous methods, EBC utilizes integer-valued bins, effectively reducing ambiguity near bin boundaries. Additionally, it incorporates a regression loss based on density maps to improve the prediction of count values. Within our backbone-agnostic EBC framework, we then introduce CLIP-EBC to fully leverage CLIP's recognition capabilities for this task. Extensive experiments demonstrate the effectiveness of EBC and the competitive performance of CLIP-EBC. Specifically, our EBC framework can improve existing classification-based methods by up to 44.5% on the UCF-QNRF dataset, and CLIP-EBC achieves state-of-the-art performance on the NWPU-Crowd test set, with an MAE of 58.2 and an RMSE of 268.5, representing improvements of 8.6% and 13.3% over the previous best method, STEERER. The code and weights are available at https://github.com/Yiming-M/CLIP-EBC.
Authors: Sierra Bonilla, Shuai Zhang, Dimitrios Psychogyios, Danail Stoyanov, Francisco Vasconcelos, Sophia Bano
Abstract: Within colorectal cancer diagnostics, conventional colonoscopy techniques face critical limitations, including a limited field of view and a lack of depth information, which can impede the detection of precancerous lesions. Current methods struggle to provide comprehensive and accurate 3D reconstructions of the colonic surface which can help minimize the missing regions and reinspection for pre-cancerous polyps. Addressing this, we introduce 'Gaussian Pancakes', a method that leverages 3D Gaussian Splatting (3D GS) combined with a Recurrent Neural Network-based Simultaneous Localization and Mapping (RNNSLAM) system. By introducing geometric and depth regularization into the 3D GS framework, our approach ensures more accurate alignment of Gaussians with the colon surface, resulting in smoother 3D reconstructions with novel viewing of detailed textures and structures. Evaluations across three diverse datasets show that Gaussian Pancakes enhances novel view synthesis quality, surpassing current leading methods with a 18% boost in PSNR and a 16% improvement in SSIM. It also delivers over 100X faster rendering and more than 10X shorter training times, making it a practical tool for real-time applications. Hence, this holds promise for achieving clinical translation for better detection and diagnosis of colorectal cancer.
Authors: Mu Hu, Wei Yin, Chi Zhang, Zhipeng Cai, Xiaoxiao Long, Hao Chen, Kaixuan Wang, Gang Yu, Chunhua Shen, Shaojie Shen
Abstract: We introduce Metric3D v2, a geometric foundation model for zero-shot metric depth and surface normal estimation from a single image, which is crucial for metric 3D recovery. While depth and normal are geometrically related and highly complimentary, they present distinct challenges. SoTA monocular depth methods achieve zero-shot generalization by learning affine-invariant depths, which cannot recover real-world metrics. Meanwhile, SoTA normal estimation methods have limited zero-shot performance due to the lack of large-scale labeled data. To tackle these issues, we propose solutions for both metric depth estimation and surface normal estimation. For metric depth estimation, we show that the key to a zero-shot single-view model lies in resolving the metric ambiguity from various camera models and large-scale data training. We propose a canonical camera space transformation module, which explicitly addresses the ambiguity problem and can be effortlessly plugged into existing monocular models. For surface normal estimation, we propose a joint depth-normal optimization module to distill diverse data knowledge from metric depth, enabling normal estimators to learn beyond normal labels. Equipped with these modules, our depth-normal models can be stably trained with over 16 million of images from thousands of camera models with different-type annotations, resulting in zero-shot generalization to in-the-wild images with unseen camera settings. Our method enables the accurate recovery of metric 3D structures on randomly collected internet images, paving the way for plausible single-image metrology. Our project page is at https://JUGGHM.github.io/Metric3Dv2.
Authors: Wenshuo Chen, Hongru Xiao, Erhang Zhang, Lijie Hu, Lei Wang, Mengyuan Liu, Chen Chen
Abstract: Is the Text to Motion model robust? Recent advancements in Text to Motion models primarily stem from more accurate predictions of specific actions. However, the text modality typically relies solely on pre-trained Contrastive Language-Image Pretraining (CLIP) models. Our research has uncovered a significant issue with the text-to-motion model: its predictions often exhibit inconsistent outputs, resulting in vastly different or even incorrect poses when presented with semantically similar or identical text inputs. In this paper, we undertake an analysis to elucidate the underlying causes of this instability, establishing a clear link between the unpredictability of model outputs and the erratic attention patterns of the text encoder module. Consequently, we introduce a formal framework aimed at addressing this issue, which we term the Stable Text-to-Motion Framework (SATO). SATO consists of three modules, each dedicated to stable attention, stable prediction, and maintaining a balance between accuracy and robustness trade-off. We present a methodology for constructing an SATO that satisfies the stability of attention and prediction. To verify the stability of the model, we introduced a new textual synonym perturbation dataset based on HumanML3D and KIT-ML. Results show that SATO is significantly more stable against synonyms and other slight perturbations while keeping its high accuracy performance.
Authors: Yuchuan Deng, Zhanpeng Hu, Jiakun Han, Chuang Deng, Qijun Zhao
Abstract: Text-based Person Re-identification (TPR) aims to retrieve specific individual images from datasets based on textual descriptions. Existing TPR methods primarily focus on recognizing explicit and positive characteristics, often overlooking the role of negative descriptions. This oversight can lead to false positives-images that meet positive criteria but should be excluded based on negative descriptions. To address these limitations, we introduce DualFocus, a unified framework that integrates plausible descriptions to enhance the interpretative accuracy of vision-language models in TPR tasks. DualFocus leverages Dual (Positive/Negative) Attribute Prompt Learning (DAPL), which incorporates Dual Image-Attribute Contrastive (DIAC) Learning and Sensitive Image-Attributes Matching (SIAM) Learning, enabling the detection of non-existent attributes and reducing false positives. To achieve a balance between coarse and fine-grained alignment of visual and textual embeddings, we propose the Dynamic Tokenwise Similarity (DTS) loss, which refines the representation of both matching and non-matching descriptions, thereby improving the matching process through detailed and adaptable similarity assessments. The comprehensive experiments on CUHK-PEDES, ICFG-PEDES, and RSTPReid, DualFocus demonstrates superior performance over state-of-the-art methods, significantly enhancing both precision and robustness in TPR.
Authors: Jianhua Zhu, Liangcai Gao, Wenqi Zhao
Abstract: Significant progress has been made in the field of handwritten mathematical expression recognition, while existing encoder-decoder methods are usually difficult to model global information in $LaTeX$. Therefore, this paper introduces a novel approach, Implicit Character-Aided Learning (ICAL), to mine the global expression information and enhance handwritten mathematical expression recognition. Specifically, we propose the Implicit Character Construction Module (ICCM) to predict implicit character sequences and use a Fusion Module to merge the outputs of the ICCM and the decoder, thereby producing corrected predictions. By modeling and utilizing implicit character information, ICAL achieves a more accurate and context-aware interpretation of handwritten mathematical expressions. Experimental results demonstrate that ICAL notably surpasses the state-of-the-art(SOTA) models, improving the expression recognition rate (ExpRate) by 2.25\%/1.81\%/1.39\% on the CROHME 2014/2016/2019 datasets respectively, and achieves a remarkable 69.06\% on the challenging HME100k test set. We make our code available on the GitHub: https://github.com/qingzhenduyu/ICAL
Authors: Wenbo Sui, Daniel Lichau, Josselin Lef\`evre, Harold Phelippeau
Abstract: Recent studies of multimodal industrial anomaly detection (IAD) based on 3D point clouds and RGB images have highlighted the importance of exploiting the redundancy and complementarity among modalities for accurate classification and segmentation. However, achieving multimodal IAD in practical production lines remains a work in progress. It is essential to consider the trade-offs between the costs and benefits associated with the introduction of new modalities while ensuring compatibility with current processes. Existing quality control processes combine rapid in-line inspections, such as optical and infrared imaging with high-resolution but time-consuming near-line characterization techniques, including industrial CT and electron microscopy to manually or semi-automatically locate and analyze defects in the production of Li-ion batteries and composite materials. Given the cost and time limitations, only a subset of the samples can be inspected by all in-line and near-line methods, and the remaining samples are only evaluated through one or two forms of in-line inspection. To fully exploit data for deep learning-driven automatic defect detection, the models must have the ability to leverage multimodal training and handle incomplete modalities during inference. In this paper, we propose CMDIAD, a Cross-Modal Distillation framework for IAD to demonstrate the feasibility of a Multi-modal Training, Few-modal Inference (MTFI) pipeline. Our findings show that the MTFI pipeline can more effectively utilize incomplete multimodal information compared to applying only a single modality for training and inference. Moreover, we investigate the reasons behind the asymmetric performance improvement using point clouds or RGB images as the main modality of inference. This provides a foundation for our future multimodal dataset construction with additional modalities from manufacturing scenarios.
Authors: Dominic LaBella, Katherine Schumacher, Michael Mix, Kevin Leu, Shan McBurney-Lin, Pierre Nedelec, Javier Villanueva-Meyer, Jonathan Shapey, Tom Vercauteren, Kazumi Chia, Omar Al-Salihi, Justin Leu, Lia Halasz, Yury Velichko, Chunhao Wang, John Kirkpatrick, Scott Floyd, Zachary J. Reitman, Trey Mullikin, Ulas Bagci, Sean Sachdev, Jona A. Hattangadi-Gluth, Tyler Seibert, Nikdokht Farid, Connor Puett, Matthew W. Pease, Kevin Shiue, Syed Muhammad Anwar, Shahriar Faghani, Muhammad Ammar Haider, Pranav Warman, Jake Albrecht, Andr\'as Jakab, Mana Moassefi, Verena Chung, Alejandro Aristizabal, Alexandros Karargyris, Hasan Kassem, Sarthak Pati, Micah Sheller, Christina Huang, Aaron Coley, Siddharth Ghanta, Alex Schneider, Conrad Sharp, Rachit Saluja, Florian Kofler, Philipp Lohmann, Phillipp Vollmuth, Louis Gagnon, Maruf Adewole, Hongwei Bran Li, Anahita Fathi Kazerooni, Nourel Hoda Tahon, Udunna Anazodo, Ahmed W. Moawad, Bjoern Menze, Marius George Linguraru, Mariam Aboian, Benedikt Wiestler, Ujjwal Baid, Gian-Marco Conte, Andreas M. Rauschecker, Ayman Nada, Aly H. Abayazeed, Raymond Huang, Maria Correia de Verdier, Jeffrey D. Rudie, Spyridon Bakas, Evan Calabrese
Abstract: The 2024 Brain Tumor Segmentation Meningioma Radiotherapy (BraTS-MEN-RT) challenge aims to advance automated segmentation algorithms using the largest known multi-institutional dataset of radiotherapy planning brain MRIs with expert-annotated target labels for patients with intact or postoperative meningioma that underwent either conventional external beam radiotherapy or stereotactic radiosurgery. Each case includes a defaced 3D post-contrast T1-weighted radiotherapy planning MRI in its native acquisition space, accompanied by a single-label "target volume" representing the gross tumor volume (GTV) and any at-risk postoperative site. Target volume annotations adhere to established radiotherapy planning protocols, ensuring consistency across cases and institutions. For preoperative meningiomas, the target volume encompasses the entire GTV and associated nodular dural tail, while for postoperative cases, it includes at-risk resection cavity margins as determined by the treating institution. Case annotations were reviewed and approved by expert neuroradiologists and radiation oncologists. Participating teams will develop, containerize, and evaluate automated segmentation models using this comprehensive dataset. Model performance will be assessed using an adapted lesion-wise Dice Similarity Coefficient and the 95% Hausdorff distance. The top-performing teams will be recognized at the Medical Image Computing and Computer Assisted Intervention Conference in October 2024. BraTS-MEN-RT is expected to significantly advance automated radiotherapy planning by enabling precise tumor segmentation and facilitating tailored treatment, ultimately improving patient outcomes.
Authors: Mohammad Erfan Sadeghi, Arash Fayyazi, Seyedarmin Azizi, Massoud Pedram
Abstract: The deployment of Vision Transformers (ViTs) on hardware platforms, specially Field-Programmable Gate Arrays (FPGAs), presents many challenges, which are mainly due to the substantial computational and power requirements of their non-linear functions, notably layer normalization, softmax, and Gaussian Error Linear Unit (GELU). These critical functions pose significant obstacles to efficient hardware implementation due to their complex mathematical operations and the inherent resource count and architectural limitations of FPGAs. PEANO-ViT offers a novel approach to streamlining the implementation of the layer normalization layer by introducing a division-free technique that simultaneously approximates the division and square root function. Additionally, PEANO-ViT provides a multi-scale division strategy to eliminate division operations in the softmax layer, aided by a Pade-based approximation for the exponential function. Finally, PEANO-ViT introduces a piece-wise linear approximation for the GELU function, carefully designed to bypass the computationally intensive operations associated with GELU. In our comprehensive evaluations, PEANO-ViT exhibits minimal accuracy degradation (<= 0.5% for DeiT-B) while significantly enhancing power efficiency, achieving improvements of 1.91x, 1.39x, 8.01x for layer normalization, softmax, and GELU, respectively. This improvement is achieved through substantial reductions in DSP, LUT, and register counts for these non-linear operations. Consequently, PEANO-ViT enables efficient deployment of Vision Transformers on resource- and power-constrained FPGA platforms.
Authors: Renqi Chen, Xinzhe Zheng, Haoyang Su, Kehan Wu
Abstract: Precise ultrasound segmentation is vital for clinicians to provide comprehensive diagnoses. However, developing a model that accurately segments ultrasound images is challenging due to the images' low quality and the scarcity of extensive labeled data. This results in two main solutions: (1) optimizing multi-scale feature representations, and (2) increasing resistance to data dependency. The first approach necessitates an advanced network architecture, but a handcrafted network is knowledge-intensive and often yields limited improvement. In contrast, neural architecture search (NAS) can more easily attain optimal performance, albeit with significant computational costs. Regarding the second issue, semi-supervised learning (SSL) is an established method, but combining it with complex NAS faces the risk of overfitting to a few labeled samples without extra constraints. Therefore, we introduce a hybrid constraint-driven semi-supervised Transformer-NAS (HCS-TNAS), balancing both solutions for segmentation. HCS-TNAS includes an Efficient NAS-ViT module for multi-scale token search before ViT's attention calculation, effectively capturing contextual and local information with lower computational costs, and a hybrid SSL framework that adds network independence and contrastive learning to the optimization for solving data dependency. By further developing a stage-wise optimization strategy, a rational network structure is identified. Experiments on public datasets show that HCS-TNAS achieves state-of-the-art performance, pushing the limit of ultrasound segmentation.
Authors: Vitor Hideyo Isume (Osaka University), Takuya Kiyokawa (Osaka University), Natsuki Yamanobe (AIST), Yukiyasu Domae (AIST), Weiwei Wan (Osaka University), Kensuke Harada (Osaka University, AIST)
Abstract: Inspired by traditional handmade crafts, where a person improvises assemblies based on the available objects, we formally introduce the Craft Assembly Task. It is a robotic assembly task that involves building an accurate representation of a given target object using the available objects, which do not directly correspond to its parts. In this work, we focus on selecting the subset of available objects for the final craft, when the given input is an RGB image of the target in the wild. We use a mask segmentation neural network to identify visible parts, followed by retrieving labelled template meshes. These meshes undergo pose optimization to determine the most suitable template. Then, we propose to simplify the parts of the transformed template mesh to primitive shapes like cuboids or cylinders. Finally, we design a search algorithm to find correspondences in the scene based on local and global proportions. We develop baselines for comparison that consider all possible combinations, and choose the highest scoring combination for common metrics used in foreground maps and mask accuracy. Our approach achieves comparable results to the baselines for two different scenes, and we show qualitative results for an implementation in a real-world scenario.
Authors: Hanjun Luo, Haoyu Huang, Ziye Deng, Xuecheng Liu, Ruizhe Chen, Zuozhu Liu
Abstract: Text-to-Image (T2I) generative models are becoming increasingly crucial due to their ability to generate high-quality images, which also raises concerns about the social biases in their outputs, especially in the human generation. Sociological research has established systematic classifications of bias. However, existing bias research about T2I models conflates different types of bias, impeding methodological progress. In this paper, we introduce BIGbench, a unified benchmark for Biases of Image Generation, featuring a meticulously designed dataset. Unlike existing benchmarks, BIGbench classifies and evaluates biases across four dimensions: manifestation of bias, visibility of bias, acquired attributes, and protected attributes, which ensures exceptional accuracy for analysis. Furthermore, BIGbench applies advanced multi-modal large language models to achieve fully automated and highly accurate evaluations. We apply BIGbench to evaluate eight representative general T2I models and three debiased methods. Our human evaluation results underscore BIGbench's effectiveness in aligning images and identifying various biases. Besides, our study also reveal new research directions about biases, such as the effect of distillation and irrelevant protected attributes. Our benchmark is openly accessible at https://github.com/BIGbench2024/BIGbench2024/ to ensure reproducibility.
Authors: Shuang Chen, Amir Atapour-Abarghouei, Haozheng Zhang, Hubert P. H. Shum
Abstract: Image inpainting, or image completion, is a crucial task in computer vision that aims to restore missing or damaged regions of images with semantically coherent content. This technique requires a precise balance of local texture replication and global contextual understanding to ensure the restored image integrates seamlessly with its surroundings. Traditional methods using Convolutional Neural Networks (CNNs) are effective at capturing local patterns but often struggle with broader contextual relationships due to the limited receptive fields. Recent advancements have incorporated transformers, leveraging their ability to understand global interactions. However, these methods face computational inefficiencies and struggle to maintain fine-grained details. To overcome these challenges, we introduce MxT composed of the proposed Hybrid Module (HM), which combines Mamba with the transformer in a synergistic manner. Mamba is adept at efficiently processing long sequences with linear computational costs, making it an ideal complement to the transformer for handling long-scale data interactions. Our HM facilitates dual-level interaction learning at both pixel and patch levels, greatly enhancing the model to reconstruct images with high quality and contextual accuracy. We evaluate MxT on the widely-used CelebA-HQ and Places2-standard datasets, where it consistently outperformed existing state-of-the-art methods. The code will be released: {\url{https://github.com/ChrisChen1023/MxT}}.
Authors: Hanjun Luo, Ziye Deng, Haoyu Huang, Xuecheng Liu, Ruizhe Chen, Zuozhu Liu
Abstract: With the rapid development of Text-to-Image (T2I) models, biases in human image generation against demographic social groups become a significant concern, impacting fairness and ethical standards in AI. Some researchers propose their methods to tackle with the issue. However, existing methods are designed for specific models with fixed prompts, limiting their adaptability to the fast-evolving models and diverse practical scenarios. Moreover, they neglect the impact of hallucinations, leading to discrepancies between expected and actual results. To address these issues, we introduce VersusDebias, a novel and universal debiasing framework for biases in arbitrary T2I models, consisting of an array generation (AG) module and an image generation (IG) module. The self-adaptive AG module generates specialized attribute arrays to post-process hallucinations and debias multiple attributes simultaneously. The IG module employs a small language model to modify prompts according to the arrays and drives the T2I model to generate debiased images, enabling zero-shot debiasing. Extensive experiments demonstrate VersusDebias's capability to debias any models across gender, race, and age simultaneously. In both zero-shot and few-shot scenarios, VersusDebias outperforms existing methods, showcasing its exceptional utility. Our work is accessible at https://github.com/VersusDebias/VersusDebias to ensure reproducibility and facilitate further research.
Authors: Lianghao Tan, Shubing Liu, Jing Gao, Xiaoyi Liu, Linyue Chu, Huangqi Jiang
Abstract: With the rapid advancement of deep learning technologies, computer vision has shown immense potential in retail automation. This paper presents a novel self-checkout system for retail based on an improved YOLOv10 network, aimed at enhancing checkout efficiency and reducing labor costs. We propose targeted optimizations to the YOLOv10 model, by incorporating the detection head structure from YOLOv8, which significantly improves product recognition accuracy. Additionally, we develop a post-processing algorithm tailored for self-checkout scenarios, to further enhance the application of system. Experimental results demonstrate that our system outperforms existing methods in both product recognition accuracy and checkout speed. This research not only provides a new technical solution for retail automation but offers valuable insights into optimizing deep learning models for real-world applications.
Authors: Yihong Lin, Zhaoxin Fan, Lingyu Xiong, Liang Peng, Xiandong Li, Wenxiong Kang, Xianjia Wu, Songju Lei, Huang Xu
Abstract: Speech-driven talking head generation is an important but challenging task for many downstream applications such as augmented reality. Existing methods have achieved remarkable performance by utilizing autoregressive models or diffusion models. However, most still suffer from modality inconsistencies, specifically the misalignment between audio and mesh modalities, which causes inconsistencies in motion diversity and lip-sync accuracy. To address this issue, this paper introduces GLDiTalker, a novel speech-driven 3D facial animation model that employs a Graph Latent Diffusion Transformer. The core idea behind GLDiTalker is that the audio-mesh modality misalignment can be resolved by diffusing the signal in a latent quantilized spatial-temporal space. To achieve this, GLDiTalker builds upon a quantilized space-time diffusion training pipeline, which consists of a Graph Enhanced Quantilized Space Learning Stage and a Space-Time Powered Latent Diffusion Stage. The first stage ensures lip-sync accuracy, while the second stage enhances motion diversity. Together, these stages enable GLDiTalker to generate temporally and spatially stable, realistic models. Extensive evaluations on several widely used benchmarks demonstrate that our method achieves superior performance compared to existing methods.
Authors: Jialang Xu, Jiacheng Wang, Lequan Yu, Danail Stoyanov, Yueming Jin, Evangelos B. Mazomenos
Abstract: Personalized federated learning (PFL) for surgical instrument segmentation (SIS) is a promising approach. It enables multiple clinical sites to collaboratively train a series of models in privacy, with each model tailored to the individual distribution of each site. Existing PFL methods rarely consider the personalization of multi-headed self-attention, and do not account for appearance diversity and instrument shape similarity, both inherent in surgical scenes. We thus propose PFedSIS, a novel PFL method with visual trait priors for SIS, incorporating global-personalized disentanglement (GPD), appearance-regulation personalized enhancement (APE), and shape-similarity global enhancement (SGE), to boost SIS performance in each site. GPD represents the first attempt at head-wise assignment for multi-headed self-attention personalization. To preserve the unique appearance representation of each site and gradually leverage the inter-site difference, APE introduces appearance regulation and provides customized layer-wise aggregation solutions via hypernetworks for each site's personalized parameters. The mutual shape information of instruments is maintained and shared via SGE, which enhances the cross-style shape consistency on the image level and computes the shape-similarity contribution of each site on the prediction level for updating the global parameters. PFedSIS outperforms state-of-the-art methods with +1.51% Dice, +2.11% IoU, -2.79 ASSD, -15.55 HD95 performance gains. The corresponding code and models will be released at https://github.com/wzjialang/PFedSIS.
Authors: Lianwei Yang, Haisong Gong, Qingyi Gu
Abstract: Vision transformers (ViTs) have garnered significant attention for their performance in vision tasks, but the high computational cost and significant latency issues have hindered widespread adoption. Post-training quantization (PTQ), a promising method for model compression, still faces accuracy degradation challenges with ViTs. There are two reasons for this: the existing quantization paradigm does not fit the power-law distribution of post-Softmax activations well, and accuracy inevitably decreases after reparameterizing post-LayerNorm activations. We propose a Distribution-Friendly and Outlier-Aware Post-training Quantization method for Vision Transformers, named DopQ-ViT. DopQ-ViT analyzes the inefficiencies of current quantizers and introduces a distribution-friendly Tan Quantizer called TanQ. TanQ focuses more on values near 1, more accurately preserving the power-law distribution of post-Softmax activations, and achieves favorable results. Besides, during the reparameterization of post-LayerNorm activations from channel-wise to layer-wise quantization, the accuracy degradation is mainly due to the significant impact of outliers in the scaling factors. Therefore, DopQ-ViT proposes a method to select Median as the Optimal Scaling Factor, denoted as MOSF, which compensates for the influence of outliers and preserves the performance of the quantization model. DopQ-ViT has been extensively validated and significantly improves the performance of quantization models, especially in low-bit settings.
Authors: Ziyang Chen, Yiwen Ye, Yongsheng Pan, Yong Xia
Abstract: Although recent years have witnessed significant advancements in medical image segmentation, the pervasive issue of domain shift among medical images from diverse centres hinders the effective deployment of pre-trained models. Many Test-time Adaptation (TTA) methods have been proposed to address this issue by fine-tuning pre-trained models with test data during inference. These methods, however, often suffer from less-satisfactory optimization due to suboptimal optimization direction (dictated by the gradient) and fixed step-size (predicated on the learning rate). In this paper, we propose the Gradient alignment-based Test-time adaptation (GraTa) method to improve both the gradient direction and learning rate in the optimization procedure. Unlike conventional TTA methods, which primarily optimize the pseudo gradient derived from a self-supervised objective, our method incorporates an auxiliary gradient with the pseudo one to facilitate gradient alignment. Such gradient alignment enables the model to excavate the similarities between different gradients and correct the gradient direction to approximate the empirical gradient related to the current segmentation task. Additionally, we design a dynamic learning rate based on the cosine similarity between the pseudo and auxiliary gradients, thereby empowering the adaptive fine-tuning of pre-trained models on diverse test data. Extensive experiments establish the effectiveness of the proposed gradient alignment and dynamic learning rate and substantiate the superiority of our GraTa method over other state-of-the-art TTA methods on a benchmark medical image segmentation task. The code and weights of pre-trained source models will be available.
Authors: Samta Rani, Tanvir Ahmad, Sarfaraz Masood, Chandni Saxena
Abstract: Despite significant research efforts and advancements, cancer remains a leading cause of mortality. Early cancer prediction has become a crucial focus in cancer research to streamline patient care and improve treatment outcomes. Manual tumor detection by histopathologists can be time consuming, prompting the need for computerized methods to expedite treatment planning. Traditional approaches to tumor detection rely on supervised learning, necessitates a large amount of annotated data for model training. However, acquiring such extensive labeled data can be laborious and time\-intensive. This research examines the three learning environments: supervised learning (SL), semi\-supervised learning (Semi\-SL), and self\-supervised learning (Self\-SL): to predict kidney, lung, and breast cancer. Three pre\-trained deep learning models (Residual Network\-50, Visual Geometry Group\-16, and EfficientNetB0) are evaluated based on these learning settings using seven carefully curated training sets. To create the first training set (TS1), SL is applied to all annotated image samples. Five training sets (TS2\-TS6) with different ratios of labeled and unlabeled cancer images are used to evaluateSemi\-SL. Unlabeled cancer images from the final training set (TS7) are utilized for Self\-SL assessment. Among different learning environments, outcomes from the Semi\-SL setting show a strong degree of agreement with the outcomes achieved in the SL setting. The uniform pattern of observations from the pre\-trained models across all three datasets validates the methodology and techniques of the research. Based on modest number of labeled samples and minimal computing cost, our study suggests that the Semi\-SL option can be a highly viable replacement for the SL option under label annotation constraint scenarios.
Authors: Aohan Li, Jiaxin Chen, Xin Liao, Dengyong Zhang
Abstract: The rapid advancement of social media platforms has significantly reduced the cost of information dissemination, yet it has also led to a proliferation of fake news, posing a threat to societal trust and credibility. Most of fake news detection research focused on integrating text and image information to represent the consistency of multiple modes in news content, while paying less attention to inconsistent information. Besides, existing methods that leveraged inconsistent information often caused one mode overshadowing another, leading to ineffective use of inconsistent clue. To address these issues, we propose an adaptive multi-modal feature fusion network (MFF-Net). Inspired by human judgment processes for determining truth and falsity in news, MFF-Net focuses on inconsistent parts when news content is generally consistent and consistent parts when it is generally inconsistent. Specifically, MFF-Net extracts semantic and global features from images and texts respectively, and learns consistency information between modes through a multiple feature fusion module. To deal with the problem of modal information being easily masked, we design a single modal feature filtering strategy to capture inconsistent information from corresponding modes separately. Finally, similarity scores are calculated based on global features with adaptive adjustments made to achieve weighted fusion of consistent and inconsistent features. Extensive experimental results demonstrate that MFF-Net outperforms state-of-the-art methods across three public news datasets derived from real social medias.
Authors: Qiming Xia, Hongwei Lin, Wei Ye, Hai Wu, Yadan Luo, Shijia Zhao, Xin Li, Chenglu Wen
Abstract: LiDAR-based outdoor 3D object detection has received widespread attention. However, training 3D detectors from the LiDAR point cloud typically relies on expensive bounding box annotations. This paper presents OC3D, an innovative weakly supervised method requiring only coarse clicks on the bird's eye view of the 3D point cloud. A key challenge here is the absence of complete geometric descriptions of the target objects from such simple click annotations. To address this problem, our proposed OC3D adopts a two-stage strategy. In the first stage, we initially design a novel dynamic and static classification strategy and then propose the Click2Box and Click2Mask modules to generate box-level and mask-level pseudo-labels for static and dynamic instances, respectively. In the second stage, we design a Mask2Box module, leveraging the learning capabilities of neural networks to update mask-level pseudo-labels, which contain less information, to box-level pseudo-labels. Experimental results on the widely used KITTI and nuScenes datasets demonstrate that our OC3D with only coarse clicks achieves state-of-the-art performance compared to weakly-supervised 3D detection methods. Combining OC3D with a missing click mining strategy, we propose an OC3D++ pipeline, which requires only 0.2% annotation cost in the KITTI dataset to achieve performance comparable to fully supervised methods. The code will be made publicly available.
Authors: Jiasong Feng, Ao Ma, Jing Wang, Bo Cheng, Xiaodan Liang, Dawei Leng, Yuhui Yin
Abstract: Synthesizing motion-rich and temporally consistent videos remains a challenge in artificial intelligence, especially when dealing with extended durations. Existing text-to-video (T2V) models commonly employ spatial cross-attention for text control, equivalently guiding different frame generations without frame-specific textual guidance. Thus, the model's capacity to comprehend the temporal logic conveyed in prompts and generate videos with coherent motion is restricted. To tackle this limitation, we introduce FancyVideo, an innovative video generator that improves the existing text-control mechanism with the well-designed Cross-frame Textual Guidance Module (CTGM). Specifically, CTGM incorporates the Temporal Information Injector (TII), Temporal Affinity Refiner (TAR), and Temporal Feature Booster (TFB) at the beginning, middle, and end of cross-attention, respectively, to achieve frame-specific textual guidance. Firstly, TII injects frame-specific information from latent features into text conditions, thereby obtaining cross-frame textual conditions. Then, TAR refines the correlation matrix between cross-frame textual conditions and latent features along the time dimension. Lastly, TFB boosts the temporal consistency of latent features. Extensive experiments comprising both quantitative and qualitative evaluations demonstrate the effectiveness of FancyVideo. Our video demo, code and model are available at https://360cvgroup.github.io/FancyVideo/.
Authors: Shuai Yuan, Hanlin Qin, Renke Kou, Xiang Yan, Zechuan Li, Chenxu Peng, Abd-Krim Seghouane
Abstract: In this work, we make the first attempt to construct a learning-based single-point annotation paradigm for infrared small target label generation (IRSTLG). Our intuition is that label generation requires just one more point prompt than target detection: IRSTLG can be regarded as an infrared small target detection (IRSTD) task with the target location hint. Based on this insight, we introduce an energy double guided single-point prompt (EDGSP) framework, which adeptly transforms the target detection network into a refined label generation method. Specifically, the proposed EDGSP includes: 1) target energy initialization (TEI) to create a foundational outline for sufficient shape evolution of pseudo label, 2) double prompt embedding (DPE) for rapid localization of interested regions and reinforcement of individual differences to avoid label adhesion, and 3) bounding box-based matching (BBM) to eliminate false alarms. Experimental results show that pseudo labels generated by three baselines equipped with EDGSP achieve 100% object-level probability of detection (Pd) and 0% false-alarm rate (Fa) on SIRST, NUDT-SIRST, and IRSTD-1k datasets, with a pixel-level intersection over union (IoU) improvement of 13.28% over state-of-the-art (SOTA) label generation methods. In the practical application of downstream IRSTD, EDGSP realizes, for the first time, a single-point generated pseudo mask beyond the full label. Even with coarse single-point annotations, it still achieves 99.5% performance of full labeling.
Authors: Abhinav Sagar
Abstract: Deep learning motivated by convolutional neural networks has been highly successful in a range of medical imaging problems like image classification, image segmentation, image synthesis etc. However for validation and interpretability, not only do we need the predictions made by the model but also how confident it is while making those predictions. This is important in safety critical applications for the people to accept it. In this work, we used an encoder decoder architecture based on variational inference techniques for segmenting brain tumour images. We evaluate our work on the publicly available BRATS dataset using Dice Similarity Coefficient (DSC) and Intersection Over Union (IOU) as the evaluation metrics. Our model is able to segment brain tumours while taking into account both aleatoric uncertainty and epistemic uncertainty in a principled bayesian manner.
Authors: Harsh G. Kamath, Manmeet Singh, Neetiraj Malviya, Alberto Martilli, Liu He, Daniel Aliaga, Cenlin He, Fei Chen, Lori A. Magruder, Zong-Liang Yang, Dev Niyogi
Abstract: We introduce University of Texas - Global Building heights for Urban Studies (UT-GLOBUS), a dataset providing building heights and urban canopy parameters (UCPs) for more than 1200 cities or locales worldwide. UT-GLOBUS combines open-source spaceborne altimetry (ICESat-2 and GEDI) and coarse-resolution urban canopy elevation data with a machine-learning model to estimate building-level information. Validation using LiDAR data from six US cities showed UT-GLOBUS-derived building heights had a root mean squared error (RMSE) of 9.1 meters. Validation of mean building heights within 1-km^2 grid cells, including data from Hamburg and Sydney, resulted in an RMSE of 7.8 meters. Testing the UCPs in the urban Weather Research and Forecasting (WRF-Urban) model resulted in a significant improvement (55% in RMSE) in intra-urban air temperature representation compared to the existing table-based local climate zone approach in Houston, TX. Additionally, we demonstrated the dataset's utility for simulating heat mitigation strategies and building energy consumption using WRF-Urban, with test cases in Chicago, IL, and Austin, TX. Street-scale mean radiant temperature simulations using the Solar and LongWave Environmental Irradiance Geometry (SOLWEIG) model, incorporating UT-GLOBUS and LiDAR-derived building heights, confirmed the dataset's effectiveness in modeling human thermal comfort in Baltimore, MD (daytime RMSE = 2.85 C). Thus, UT-GLOBUS can be used for modeling urban hazards with significant socioeconomic and biometeorological risks, enabling finer scale urban climate simulations and overcoming previous limitations due to the lack of building information.
Authors: Jay J. Yoo, Khashayar Namdar, Matthias W. Wagner, Liana Nobre, Uri Tabori, Cynthia Hawkins, Birgit B. Ertl-Wagner, Farzad Khalvati
Abstract: Segmentation of regions of interest (ROIs) for identifying abnormalities is a leading problem in medical imaging. Using machine learning for this problem generally requires manually annotated ground-truth segmentations, demanding extensive time and resources from radiologists. This work presents a weakly supervised approach that utilizes binary image-level labels, which are much simpler to acquire, to effectively segment anomalies in 2D magnetic resonance images without ground truth annotations. We train a generative adversarial network (GAN) that converts cancerous images to healthy variants, which are used along with localization seeds as priors to generate improved weakly supervised segmentations. The non-cancerous variants can also be used to evaluate the segmentations in a weakly supervised fashion, which allows for the most effective segmentations to be identified and then applied to downstream clinical classification tasks. On the Multimodal Brain Tumor Segmentation (BraTS) 2020 dataset, our proposed method generates and identifies segmentations that achieve test Dice coefficients of 83.91%. Using these segmentations for pathology classification results with a test AUC of 93.32% which is comparable to the test AUC of 95.80% achieved when using true segmentations.
Authors: Tom\'a\v{s} Dlask, Bogdan Savchynskyy
Abstract: We study the set of optimal solutions of the dual linear programming formulation of the linear assignment problem (LAP) to propose a method for computing a solution from the relative interior of this set. Assuming that an arbitrary dual-optimal solution and an optimal assignment are available (for which many efficient algorithms already exist), our method computes a relative-interior solution in linear time. Since the LAP occurs as a subproblem in the linear programming (LP) relaxation of the quadratic assignment problem (QAP), we employ our method as a new component in the family of dual-ascent algorithms that provide bounds on the optimal value of the QAP. To make our results applicable to the incomplete QAP, which is of interest in practical use-cases, we also provide a linear-time reduction from the incomplete LAP to the complete LAP along with a mapping that preserves optimality and membership in the relative interior. Our experiments on publicly available benchmarks indicate that our approach with relative-interior solution can frequently provide bounds near the optimum of the LP relaxation and its runtime is much lower when compared to a commercial LP solver.
Authors: Jiazhen Pan, Wenqi Huang, Daniel Rueckert, Thomas K\"ustner, Kerstin Hammernik
Abstract: In cardiac CINE, motion-compensated MR reconstruction (MCMR) is an effective approach to address highly undersampled acquisitions by incorporating motion information between frames. In this work, we propose a novel perspective for addressing the MCMR problem and a more integrated and efficient solution to the MCMR field. Contrary to state-of-the-art (SOTA) MCMR methods which break the original problem into two sub-optimization problems, i.e. motion estimation and reconstruction, we formulate this problem as a single entity with one single optimization. Our approach is unique in that the motion estimation is directly driven by the ultimate goal, reconstruction, but not by the canonical motion-warping loss (similarity measurement between motion-warped images and target images). We align the objectives of motion estimation and reconstruction, eliminating the drawbacks of artifacts-affected motion estimation and therefore error-propagated reconstruction. Further, we can deliver high-quality reconstruction and realistic motion without applying any regularization/smoothness loss terms, circumventing the non-trivial weighting factor tuning. We evaluate our method on two datasets: 1) an in-house acquired 2D CINE dataset for the retrospective study and 2) the public OCMR cardiac dataset for the prospective study. The conducted experiments indicate that the proposed MCMR framework can deliver artifact-free motion estimation and high-quality MR images even for imaging accelerations up to 20x, outperforming SOTA non-MCMR and MCMR methods in both qualitative and quantitative evaluation across all experiments. The code is available at https://github.com/JZPeterPan/MCMR-Recon-Driven-Motion.
URLs: https://github.com/JZPeterPan/MCMR-Recon-Driven-Motion.
Authors: Byeongjoo Ahn, Karren Yang, Brian Hamilton, Jonathan Sheaffer, Anurag Ranjan, Miguel Sarabia, Oncel Tuzel, Jen-Hao Rick Chang
Abstract: We investigate the benefit of combining blind audio recordings with 3D scene information for novel-view acoustic synthesis. Given audio recordings from 2-4 microphones and the 3D geometry and material of a scene containing multiple unknown sound sources, we estimate the sound anywhere in the scene. We identify the main challenges of novel-view acoustic synthesis as sound source localization, separation, and dereverberation. While naively training an end-to-end network fails to produce high-quality results, we show that incorporating room impulse responses (RIRs) derived from 3D reconstructed rooms enables the same network to jointly tackle these tasks. Our method outperforms existing methods designed for the individual tasks, demonstrating its effectiveness at utilizing 3D visual information. In a simulated study on the Matterport3D-NVAS dataset, our model achieves near-perfect accuracy on source localization, a PSNR of 26.44dB and a SDR of 14.23dB for source separation and dereverberation, resulting in a PSNR of 25.55 dB and a SDR of 14.20 dB on novel-view acoustic synthesis. We release our code and model on our project website at https://github.com/apple/ml-nvas3d. Please wear headphones when listening to the results.
Authors: Wenxuan Yang, Weimin Tan, Yuqi Sun, Bo Yan
Abstract: Foundation models, pre-trained on massive datasets, have achieved unprecedented generalizability. However, is it truly necessary to involve such vast amounts of data in pre-training, consuming extensive computational resources? This paper introduces data-effective learning, aiming to use data in the most impactful way to pre-train foundation models. This involves strategies that focus on data quality rather than quantity, ensuring the data used for training has high informational value. Data-effective learning plays a profound role in accelerating foundation model training, reducing computational costs, and saving data storage, which is very important as the volume of medical data in recent years has grown beyond many people's expectations. However, due to the lack of standards and comprehensive benchmarks, research on medical data-effective learning is poorly studied. To address this gap, our paper introduces a comprehensive benchmark specifically for evaluating data-effective learning in the medical field. This benchmark includes a dataset with millions of data samples from 31 medical centers (DataDEL), a baseline method for comparison (MedDEL), and a new evaluation metric (NormDEL) to objectively measure data-effective learning performance. Our extensive experimental results show the baseline MedDEL can achieve performance comparable to the original large dataset with only 5% of the data. Establishing such an open data-effective learning benchmark is crucial for the medical foundation model research community because it facilitates efficient data use, promotes collaborative breakthroughs, and fosters the development of cost-effective, scalable, and impactful healthcare solutions.
Authors: Michael Potter, Murat Akcakaya, Marius Necsoiu, Gunar Schirner, Deniz Erdogmus, Tales Imbiriba
Abstract: Radar Automated Target Recognition (RATR) for Unmanned Aerial Vehicles (UAVs) involves transmitting Electromagnetic Waves (EMWs) and performing target type recognition on the received radar echo, crucial for defense and aerospace applications. Previous studies highlighted the advantages of multistatic radar configurations over monostatic ones in RATR. However, fusion methods in multistatic radar configurations often suboptimally combine classification vectors from individual radars probabilistically. To address this, we propose a fully Bayesian RATR framework employing Optimal Bayesian Fusion (OBF) to aggregate classification probability vectors from multiple radars. OBF, based on expected 0-1 loss, updates a Recursive Bayesian Classification (RBC) posterior distribution for target UAV type, conditioned on historical observations across multiple time steps. We evaluate the approach using simulated random walk trajectories for seven drones, correlating target aspect angles to Radar Cross Section (RCS) measurements in an anechoic chamber. Comparing against single radar Automated Target Recognition (ATR) systems and suboptimal fusion methods, our empirical results demonstrate that the OBF method integrated with RBC significantly enhances classification accuracy compared to other fusion methods and single radar configurations.
Authors: Sungmin Cha, Kyunghyun Cho
Abstract: In this paper, we argue that the conventional evaluation protocol in continual learning (CL) research deviates from the fundamental principle in machine learning evaluation. The primary objective of CL algorithm is to balance the trade-off between plasticity (learning new knowledge from new tasks) and stability (retaining knowledge from previous tasks). To evaluate it, a CL scenario is constructed by using a benchmark dataset, where a neural network model is continually trained on the training data of each task, and the best hyperparameters for a CL algorithm are selected based on validation data.The final evaluation involves assessing the model trained with these hyperparameters on the test data from the same scenario. This evaluation protocol primarily aims to assess how well a CL algorithm performs on unseen data within that specific scenario. However, to accurately evaluate the CL algorithm, the focus should be on assessing generalizability of each algorithm's CL capacity to handle unseen scenarios. To achieve this evaluation goal, we propose a revised evaluation protocol. Our protocol consists of two phases: hyperparameter tuning and evaluation. Both phases share the same scenario configuration (e.g., the number of tasks) but the scenarios for each phase are generated from different datasets. During the hyperparameter tuning phase, the best hyperparameters are identified, which are then used to train the model using the CL algorithm in the evaluation phase. Finally, the result from this phase is reported as the final evaluation. We apply the proposed evaluation protocol to class-incremental learning algorithms, both with and without a pretrained model. Through extensive experiments involving approximately 5000 trials, we demonstrate that most state-of-the-art algorithms fail to exhibit the reported performance, revealing a lack of generalizability.
Authors: Aiden Swann, Matthew Strong, Won Kyung Do, Gadiel Sznaier Camps, Mac Schwager, Monroe Kennedy III
Abstract: In this work, we propose a novel method to supervise 3D Gaussian Splatting (3DGS) scenes using optical tactile sensors. Optical tactile sensors have become widespread in their use in robotics for manipulation and object representation; however, raw optical tactile sensor data is unsuitable to directly supervise a 3DGS scene. Our representation leverages a Gaussian Process Implicit Surface to implicitly represent the object, combining many touches into a unified representation with uncertainty. We merge this model with a monocular depth estimation network, which is aligned in a two stage process, coarsely aligning with a depth camera and then finely adjusting to match our touch data. For every training image, our method produces a corresponding fused depth and uncertainty map. Utilizing this additional information, we propose a new loss function, variance weighted depth supervised loss, for training the 3DGS scene model. We leverage the DenseTact optical tactile sensor and RealSense RGB-D camera to show that combining touch and vision in this manner leads to quantitatively and qualitatively better results than vision or touch alone in a few-view scene syntheses on opaque as well as on reflective and transparent objects. Please see our project page at http://armlabstanford.github.io/touch-gs
Authors: Chenhui Xu, Fuxun Yu, Zirui Xu, Nathan Inkawhich, Xiang Chen
Abstract: Recent research underscores the pivotal role of the Out-of-Distribution (OOD) feature representation field scale in determining the efficacy of models in OOD detection. Consequently, the adoption of model ensembles has emerged as a prominent strategy to augment this feature representation field, capitalizing on anticipated model diversity. However, our introduction of novel qualitative and quantitative model ensemble evaluation methods, specifically Loss Basin/Barrier Visualization and the Self-Coupling Index, reveals a critical drawback in existing ensemble methods. We find that these methods incorporate weights that are affine-transformable, exhibiting limited variability and thus failing to achieve the desired diversity in feature representation. To address this limitation, we elevate the dimensions of traditional model ensembles, incorporating various factors such as different weight initializations, data holdout, etc., into distinct supervision tasks. This innovative approach, termed Multi-Comprehension (MC) Ensemble, leverages diverse training tasks to generate distinct comprehensions of the data and labels, thereby extending the feature representation field. Our experimental results demonstrate the superior performance of the MC Ensemble strategy in OOD detection compared to both the naive Deep Ensemble method and a standalone model of comparable size. This underscores the effectiveness of our proposed approach in enhancing the model's capability to detect instances outside its training distribution.
Authors: Tianhang Nan, Hao Quan, Yong Ding, Xingyu Li, Kai Yang, Xiaoyu Cui
Abstract: Multiple Instance Learning (MIL) has garnered widespread attention in the field of Whole Slide Image (WSI) classification as it replaces pixel-level manual annotation with diagnostic reports as labels, significantly reducing labor costs. Recent research has shown that bag-level MIL methods often yield better results because they can consider all patches of the WSI as a whole. However, a drawback of such methods is the incorporation of more redundant patches, leading to interference. To extract patches with high diagnostic value while excluding interfering patches to address this issue, we developed an attention-based feature distillation multi-instance learning (AFD-MIL) approach. This approach proposed the exclusion of redundant patches as a preprocessing operation in weakly supervised learning, directly mitigating interference from extensive noise. It also pioneers the use of attention mechanisms to distill features with high diagnostic value, as opposed to the traditional practice of indiscriminately and forcibly integrating all patches. Additionally, we introduced global loss optimization to finely control the feature distillation module. AFD-MIL is orthogonal to many existing MIL methods, leading to consistent performance improvements. This approach has surpassed the current state-of-the-art method, achieving 91.47% ACC (accuracy) and 94.29% AUC (area under the curve) on the Camelyon16 (Camelyon Challenge 2016, breast cancer), while 93.33% ACC and 98.17% AUC on the TCGA-NSCLC (The Cancer Genome Atlas Program: non-small cell lung cancer). Different feature distillation methods were used for the two datasets, tailored to the specific diseases, thereby improving performance and interpretability.
Authors: Jannatul Ferdaus, Sameera Pisupati, Mahedi Hasan, Sathwick Paladugu
Abstract: Federated Learning enables robots to learn from each other's experiences without relying on centralized data collection. Each robot independently maintains a model of crop conditions and chemical spray effectiveness, which is periodically shared with other robots in the fleet. A communication protocol is designed to optimize chemical spray applications by facilitating the exchange of information about crop conditions, weather, and other critical factors. The federated learning algorithm leverages this shared data to continuously refine the chemical spray strategy, reducing waste and improving crop yields. This approach has the potential to revolutionize the agriculture industry by offering a scalable and efficient solution for crop protection. However, significant challenges remain, including the development of a secure and robust communication protocol, the design of a federated learning algorithm that effectively integrates data from multiple sources, and ensuring the safety and reliability of autonomous robots. The proposed cluster-based federated learning approach also effectively reduces the computational load on the global server and minimizes communication overhead among clients.
Authors: Atefe Hassani, Islem Rekik
Abstract: A fundamental challenge in federated learning lies in mixing heterogeneous datasets and classification tasks while minimizing the high communication cost caused by clients as well as the exchange of weight updates with the server over a fixed number of rounds. This results in divergent model convergence rates and performance, which may hinder their deployment in precision medicine. In real-world scenarios, client data is collected from different hospitals with extremely varying components (e.g., imaging modality, organ type, etc). Previous studies often overlooked the convoluted heterogeneity during the training stage where the target learning tasks vary across clients as well as the dataset type and their distributions. To address such limitations, we unprecedentedly introduce UniFed, a universal federated learning paradigm that aims to classify any disease from any imaging modality. UniFed also handles the issue of varying convergence times in the client-specific optimization based on the complexity of their learning tasks. Specifically, by dynamically adjusting both local and global models, UniFed considers the varying task complexities of clients and the server, enhancing its adaptability to real-world scenarios, thereby mitigating issues related to overtraining and excessive communication. Furthermore, our framework incorporates a sequential model transfer mechanism that takes into account the diverse tasks among hospitals and a dynamic task-complexity based ordering. We demonstrate the superiority of our framework in terms of accuracy, communication cost, and convergence time over relevant benchmarks in diagnosing retina, histopathology, and liver tumour diseases under federated learning. Our UniFed code is available at https://github.com/basiralab/UniFed.
Authors: Junxian Li, Di Zhang, Xunzhi Wang, Zeying Hao, Jingdi Lei, Qian Tan, Cai Zhou, Wei Liu, Yaotian Yang, Xinrui Xiong, Weiyun Wang, Zhe Chen, Wenhai Wang, Wei Li, Shufei Zhang, Mao Su, Wanli Ouyang, Yuqiang Li, Dongzhan Zhou
Abstract: Large Language Models (LLMs) have achieved remarkable success and have been applied across various scientific fields, including chemistry. However, many chemical tasks require the processing of visual information, which cannot be successfully handled by existing chemical LLMs. This brings a growing need for models capable of integrating multimodal information in the chemical domain. In this paper, we introduce \textbf{ChemVLM}, an open-source chemical multimodal large language model specifically designed for chemical applications. ChemVLM is trained on a carefully curated bilingual multimodal dataset that enhances its ability to understand both textual and visual chemical information, including molecular structures, reactions, and chemistry examination questions. We develop three datasets for comprehensive evaluation, tailored to Chemical Optical Character Recognition (OCR), Multimodal Chemical Reasoning (MMCR), and Multimodal Molecule Understanding tasks. We benchmark ChemVLM against a range of open-source and proprietary multimodal large language models on various tasks. Experimental results demonstrate that ChemVLM achieves competitive performance across all evaluated tasks. Our model can be found at https://huggingface.co/AI4Chem/ChemVLM-26B.