Authors: Lev Ayzenberg, Raja Giryes, Hayit Greenspan
Abstract: Deep learning models have emerged as the cornerstone of medical image segmentation, but their efficacy hinges on the availability of extensive manually labeled datasets and their adaptability to unforeseen categories remains a challenge. Few-shot segmentation (FSS) offers a promising solution by endowing models with the capacity to learn novel classes from limited labeled examples. A leading method for FSS is ALPNet, which compares features between the query image and the few available support segmented images. A key question about using ALPNet is how to design its features. In this work, we delve into the potential of using features from DINOv2, which is a foundational self-supervised learning model in computer vision. Leveraging the strengths of ALPNet and harnessing the feature extraction capabilities of DINOv2, we present a novel approach to few-shot segmentation that not only enhances performance but also paves the way for more robust and adaptable medical image analysis.
Authors: Katia Jodogne-Del Litto, Guillaume-Alexandre Bilodeau
Abstract: Increasing the accuracy of instance segmentation methods is often done at the expense of speed. Using coarser representations, we can reduce the number of parameters and thus obtain real-time masks. In this paper, we take inspiration from the set cover problem to predict mask approximations. Given ground-truth binary masks of objects of interest as training input, our method learns to predict the approximate coverage of these objects by disks without supervision on their location or radius. Each object is represented by a fixed number of disks with different radii. In the learning phase, we consider the radius as proportional to a standard deviation in order to compute the error to propagate on a set of two-dimensional Gaussian functions rather than disks. We trained and tested our instance segmentation method on challenging datasets showing dense urban settings with various road users. Our method achieve state-of-the art results on the IDD and KITTI dataset with an inference time of 0.040 s on a single RTX 3090 GPU.
Authors: Sagi Eppel, Jolina Li, Manuel Drehwald, Alan Aspuru-Guzik
Abstract: Visual understanding and segmentation of materials and their states is fundamental for understanding the physical world. The infinite textures, shapes and often blurry boundaries formed by material make this task particularly hard to generalize. Whether it's identifying wet regions of a surface, minerals in rocks, infected regions in plants, or pollution in water, each material state has its own unique form. For neural nets to learn class-agnostic materials segmentation it is necessary to first collect and annotate data that capture this complexity. Collecting real-world images and manually annotating is limited both by the cost and limited precision of manual labor. In contrast, synthetic data is highly accurate and almost cost-free but fails to replicate the vast diversity of the material world. In this work, we suggest a method to bridge this crucial gap, by implanting patterns extracted from real-world images, in synthetic data. Hence, patterns automatically collected from natural images are used to map materials into synthetic scenes. This unsupervised approach allows the generated data to capture the vast complexity of the real world while maintaining the precision and scale of synthetic data. We also present the first general benchmark for class-agnostic material state segmentation. The benchmark images contain a wide range of real-world images of material states, from cooking, food, rocks, construction, plants, and liquids each in various states (wet/dry/stained/cooked/burned/worned/rusted/sediment/foam...). The annotation includes both partial similarity between regions with similar but not identical materials, and hard segmentation of only points of the exact same material state. We show that net trains on MatSeg significantly outperform existing state-of-the-art methods on this task.
Authors: Yuan Gao, Kunyu Shi, Pengkai Zhu, Edouard Belval, Oren Nuriel, Srikar Appalaraju, Shabnam Ghadar, Vijay Mahadevan, Zhuowen Tu, Stefano Soatto
Abstract: We propose Strongly Supervised pre-training with ScreenShots (S4) - a novel pre-training paradigm for Vision-Language Models using data from large-scale web screenshot rendering. Using web screenshots unlocks a treasure trove of visual and textual cues that are not present in using image-text pairs. In S4, we leverage the inherent tree-structured hierarchy of HTML elements and the spatial localization to carefully design 10 pre-training tasks with large scale annotated data. These tasks resemble downstream tasks across different domains and the annotations are cheap to obtain. We demonstrate that, compared to current screenshot pre-training objectives, our innovative pre-training method significantly enhances performance of image-to-text model in nine varied and popular downstream tasks - up to 76.1% improvements on Table Detection, and at least 1% on Widget Captioning.
Authors: Changan Chen, Rui Wang, Christoph Vogel, Marc Pollefeys
Abstract: In this paper we propose an efficient data-driven solution to self-localization within a floorplan. Floorplan data is readily available, long-term persistent and inherently robust to changes in the visual appearance. Our method does not require retraining per map and location or demand a large database of images of the area of interest. We propose a novel probabilistic model consisting of an observation and a novel temporal filtering module. Operating internally with an efficient ray-based representation, the observation module consists of a single and a multiview module to predict horizontal depth from images and fuses their results to benefit from advantages offered by either methodology. Our method operates on conventional consumer hardware and overcomes a common limitation of competing methods that often demand upright images. Our full system meets real-time requirements, while outperforming the state-of-the-art by a significant margin.
Authors: Jiajia Li, Dong Chen, Xunyuan Yin, Zhaojian Li
Abstract: Effective weed control plays a crucial role in optimizing crop yield and enhancing agricultural product quality. However, the reliance on herbicide application not only poses a critical threat to the environment but also promotes the emergence of resistant weeds. Fortunately, recent advances in precision weed management enabled by ML and DL provide a sustainable alternative. Despite great progress, existing algorithms are mainly developed based on supervised learning approaches, which typically demand large-scale datasets with manual-labeled annotations, which is time-consuming and labor-intensive. As such, label-efficient learning methods, especially semi-supervised learning, have gained increased attention in the broader domain of computer vision and have demonstrated promising performance. These methods aim to utilize a small number of labeled data samples along with a great number of unlabeled samples to develop high-performing models comparable to the supervised learning counterpart trained on a large amount of labeled data samples. In this study, we assess the effectiveness of a semi-supervised learning framework for multi-class weed detection, employing two well-known object detection frameworks, namely FCOS and Faster-RCNN. Specifically, we evaluate a generalized student-teacher framework with an improved pseudo-label generation module to produce reliable pseudo-labels for the unlabeled data. To enhance generalization, an ensemble student network is employed to facilitate the training process. Experimental results show that the proposed approach is able to achieve approximately 76\% and 96\% detection accuracy as the supervised methods with only 10\% of labeled data in CottenWeedDet3 and CottonWeedDet12, respectively. We offer access to the source code, contributing a valuable resource for ongoing semi-supervised learning research in weed detection and beyond.
Authors: Yong Li, Shiguang Shan
Abstract: Facial action unit (AU) detection, aiming to classify AU present in the facial image, has long suffered from insufficient AU annotations. In this paper, we aim to mitigate this data scarcity issue by learning AU representations from a large number of unlabelled facial videos in a contrastive learning paradigm. We formulate the self-supervised AU representation learning signals in two-fold: (1) AU representation should be frame-wisely discriminative within a short video clip; (2) Facial frames sampled from different identities but show analogous facial AUs should have consistent AU representations. As to achieve these goals, we propose to contrastively learn the AU representation within a video clip and devise a cross-identity reconstruction mechanism to learn the person-independent representations. Specially, we adopt a margin-based temporal contrastive learning paradigm to perceive the temporal AU coherence and evolution characteristics within a clip that consists of consecutive input facial frames. Moreover, the cross-identity reconstruction mechanism facilitates pushing the faces from different identities but show analogous AUs close in the latent embedding space. Experimental results on three public AU datasets demonstrate that the learned AU representation is discriminative for AU detection. Our method outperforms other contrastive learning methods and significantly closes the performance gap between the self-supervised and supervised AU detection approaches.
Authors: Liuyi Wang, Zongtao He, Ronghao Dang, Huiyi Chen, Chengju Liu, Qijun Chen
Abstract: Vision-and-Language Navigation (VLN) has gained significant research interest in recent years due to its potential applications in real-world scenarios. However, existing VLN methods struggle with the issue of spurious associations, resulting in poor generalization with a significant performance gap between seen and unseen environments. In this paper, we tackle this challenge by proposing a unified framework CausalVLN based on the causal learning paradigm to train a robust navigator capable of learning unbiased feature representations. Specifically, we establish reasonable assumptions about confounders for vision and language in VLN using the structured causal model (SCM). Building upon this, we propose an iterative backdoor-based representation learning (IBRL) method that allows for the adaptive and effective intervention on confounders. Furthermore, we introduce the visual and linguistic backdoor causal encoders to enable unbiased feature expression for multi-modalities during training and validation, enhancing the agent's capability to generalize across different environments. Experiments on three VLN datasets (R2R, RxR, and REVERIE) showcase the superiority of our proposed method over previous state-of-the-art approaches. Moreover, detailed visualization analysis demonstrates the effectiveness of CausalVLN in significantly narrowing down the performance gap between seen and unseen environments, underscoring its strong generalization capability.
Authors: Sungho Kang, YeongHyeon Park, Hyunkyu Park, Juneho Yi
Abstract: Scene depth estimation from paintings can streamline the process of 3D sculpture creation so that visually impaired people appreciate the paintings with tactile sense. However, measuring depth of oriental landscape painting images is extremely challenging due to its unique method of depicting depth and poor preservation. To address the problem of scene depth estimation from oriental landscape painting images, we propose a novel framework that consists of two-step Image-to-Image translation method with CLIP-based image matching at the front end to predict the real scene image that best matches with the given oriental landscape painting image. Then, we employ a pre-trained SOTA depth estimation model for the generated real scene image. In the first step, CycleGAN converts an oriental landscape painting image into a pseudo-real scene image. We utilize CLIP to semantically match landscape photo images with an oriental landscape painting image for training CycleGAN in an unsupervised manner. Then, the pseudo-real scene image and oriental landscape painting image are fed into DiffuseIT to predict a final real scene image in the second step. Finally, we measure depth of the generated real scene image using a pre-trained depth estimation model such as MiDaS. Experimental results show that our approach performs well enough to predict real scene images corresponding to oriental landscape painting images. To the best of our knowledge, this is the first study to measure the depth of oriental landscape painting images. Our research potentially assists visually impaired people in experiencing paintings in diverse ways. We will release our code and resulting dataset.
Authors: Sanqing Qu, Tianpei Zou, Lianghua He, Florian R\"ohrbein, Alois Knoll, Guang Chen, Changjun Jiang
Abstract: Universal Domain Adaptation (UniDA) targets knowledge transfer in the presence of both covariate and label shifts. Recently, Source-free Universal Domain Adaptation (SF-UniDA) has emerged to achieve UniDA without access to source data, which tends to be more practical due to data protection policies. The main challenge lies in determining whether covariate-shifted samples belong to target-private unknown categories. Existing methods tackle this either through hand-crafted thresholding or by developing time-consuming iterative clustering strategies. In this paper, we propose a new idea of LEArning Decomposition (LEAD), which decouples features into source-known and -unknown components to identify target-private data. Technically, LEAD initially leverages the orthogonal decomposition analysis for feature decomposition. Then, LEAD builds instance-level decision boundaries to adaptively identify target-private data. Extensive experiments across various UniDA scenarios have demonstrated the effectiveness and superiority of LEAD. Notably, in the OPDA scenario on VisDA dataset, LEAD outperforms GLC by 3.5% overall H-score and reduces 75% time to derive pseudo-labeling decision boundaries. Besides, LEAD is also appealing in that it is complementary to most existing methods. The code is available at https://github.com/ispc-lab/LEAD.
Authors: Bingyan Liu, Chengyu Wang, Tingfeng Cao, Kui Jia, Jun Huang
Abstract: Deep Text-to-Image Synthesis (TIS) models such as Stable Diffusion have recently gained significant popularity for creative Text-to-image generation. Yet, for domain-specific scenarios, tuning-free Text-guided Image Editing (TIE) is of greater importance for application developers, which modify objects or object properties in images by manipulating feature components in attention layers during the generation process. However, little is known about what semantic meanings these attention layers have learned and which parts of the attention maps contribute to the success of image editing. In this paper, we conduct an in-depth probing analysis and demonstrate that cross-attention maps in Stable Diffusion often contain object attribution information that can result in editing failures. In contrast, self-attention maps play a crucial role in preserving the geometric and shape details of the source image during the transformation to the target image. Our analysis offers valuable insights into understanding cross and self-attention maps in diffusion models. Moreover, based on our findings, we simplify popular image editing methods and propose a more straightforward yet more stable and efficient tuning-free procedure that only modifies self-attention maps of the specified attention layers during the denoising process. Experimental results show that our simplified method consistently surpasses the performance of popular approaches on multiple datasets.
Authors: Gangwei Xu, Yujin Wang, Jinwei Gu, Tianfan Xue, Xin Yang
Abstract: Reconstructing High Dynamic Range (HDR) video from image sequences captured with alternating exposures is challenging, especially in the presence of large camera or object motion. Existing methods typically align low dynamic range sequences using optical flow or attention mechanism for deghosting. However, they often struggle to handle large complex motions and are computationally expensive. To address these challenges, we propose a robust and efficient flow estimator tailored for real-time HDR video reconstruction, named HDRFlow. HDRFlow has three novel designs: an HDR-domain alignment loss (HALoss), an efficient flow network with a multi-size large kernel (MLK), and a new HDR flow training scheme. The HALoss supervises our flow network to learn an HDR-oriented flow for accurate alignment in saturated and dark regions. The MLK can effectively model large motions at a negligible cost. In addition, we incorporate synthetic data, Sintel, into our training dataset, utilizing both its provided forward flow and backward flow generated by us to supervise our flow network, enhancing our performance in large motion regions. Extensive experiments demonstrate that our HDRFlow outperforms previous methods on standard benchmarks. To the best of our knowledge, HDRFlow is the first real-time HDR video reconstruction method for video sequences captured with alternating exposures, capable of processing 720p resolution inputs at 25ms.
Authors: Ruizhuo Song, Beiming Yuan
Abstract: This paper presents significant advancements in the field of abstract reasoning, particularly for Raven's Progressive Matrices (RPM) and Bongard-Logo problems. We first introduce D2C, a method that redefines concept boundaries in these domains and bridges the gap between high-level concepts and their low-dimensional representations. Leveraging this foundation, we propose D3C, a novel approach for tackling Bongard-Logo problems. D3C estimates the distributions of image representations and measures their Sinkhorn distance to achieve remarkable reasoning accuracy. This innovative method provides new insights into the relationships between images and advances the state-of-the-art in abstract reasoning. To further enhance computational efficiency without sacrificing performance, we introduce D3C-cos. This variant of D3C constrains distribution distances, offering a more computationally efficient solution for RPM problems while maintaining high accuracy. Additionally, we present Lico-Net, a baseline network for RPM that integrates D3C and D3C-cos. By estimating and constraining the distributions of regularity representations, Lico-Net addresses both problem-solving and interpretability challenges, achieving state-of-the-art performance. Finally, we extend our methodology with D4C, an adversarial approach that further refines concept boundaries compared to D2C. Tailored for RPM and Bongard-Logo problems, D4C demonstrates significant improvements in addressing the challenges of abstract reasoning. Overall, our contributions advance the field of abstract reasoning, providing new perspectives and practical solutions to long-standing problems.
Authors: Xiangquan Gui, Binxuan Zhang, Li Li, Yi Yang
Abstract: Chinese landscape painting has a unique and artistic style, and its drawing technique is highly abstract in both the use of color and the realistic representation of objects. Previous methods focus on transferring from modern photos to ancient ink paintings. However, little attention has been paid to translating landscape paintings into modern photos. To solve such problems, in this paper, we (1) propose DLP-GAN (\textbf{D}raw Modern Chinese \textbf{L}andscape \textbf{P}hotos with \textbf{G}enerative \textbf{A}dversarial \textbf{N}etwork), an unsupervised cross-domain image translation framework with a novel asymmetric cycle mapping, and (2) introduce a generator based on a dense-fusion module to match different translation directions. Moreover, a dual-consistency loss is proposed to balance the realism and abstraction of model painting. In this way, our model can draw landscape photos and sketches in the modern sense. Finally, based on our collection of modern landscape and sketch datasets, we compare the images generated by our model with other benchmarks. Extensive experiments including user studies show that our model outperforms state-of-the-art methods.
Authors: Shanka Subhra Mondal, Jonathan D. Cohen, Taylor W. Webb
Abstract: Abstract visual reasoning is a characteristically human ability, allowing the identification of relational patterns that are abstracted away from object features, and the systematic generalization of those patterns to unseen problems. Recent work has demonstrated strong systematic generalization in visual reasoning tasks involving multi-object inputs, through the integration of slot-based methods used for extracting object-centric representations coupled with strong inductive biases for relational abstraction. However, this approach was limited to problems containing a single rule, and was not scalable to visual reasoning problems containing a large number of objects. Other recent work proposed Abstractors, an extension of Transformers that incorporates strong relational inductive biases, thereby inheriting the Transformer's scalability and multi-head architecture, but it has yet to be demonstrated how this approach might be applied to multi-object visual inputs. Here we combine the strengths of the above approaches and propose Slot Abstractors, an approach to abstract visual reasoning that can be scaled to problems involving a large number of objects and multiple relations among them. The approach displays state-of-the-art performance across four abstract visual reasoning tasks.
Authors: Cheng-Yen Yang, Hsiang-Wei Huang, Zhongyu Jiang, Hao Wang, Farron Wallace, Jenq-Neng Hwang
Abstract: Dense object counting or crowd counting has come a long way thanks to the recent development in the vision community. However, indiscernible object counting, which aims to count the number of targets that are blended with respect to their surroundings, has been a challenge. Image-based object counting datasets have been the mainstream of the current publicly available datasets. Therefore, we propose a large-scale dataset called YoutubeFish-35, which contains a total of 35 sequences of high-definition videos with high frame-per-second and more than 150,000 annotated center points across a selected variety of scenes. For benchmarking purposes, we select three mainstream methods for dense object counting and carefully evaluate them on the newly collected dataset. We propose TransVidCount, a new strong baseline that combines density and regression branches along the temporal domain in a unified framework and can effectively tackle indiscernible object counting with state-of-the-art performance on YoutubeFish-35 dataset.
Authors: Hao Wang, Sayed Pedram Haeri Boroujeni, Xiwen Chen, Ashish Bastola, Huayu Li, Abolfazl Razi
Abstract: The rise of machine learning in recent years has brought benefits to various research fields such as wide fire detection. Nevertheless, small object detection and rare object detection remain a challenge. To address this problem, we present a dataset automata that can generate ground truth paired datasets using diffusion models. Specifically, we introduce a mask-guided diffusion framework that can fusion the wildfire into the existing images while the flame position and size can be precisely controlled. In advance, to fill the gap that the dataset of wildfire images in specific scenarios is missing, we vary the background of synthesized images by controlling both the text prompt and input image. Furthermore, to solve the color tint problem or the well-known domain shift issue, we apply the CLIP model to filter the generated massive dataset to preserve quality. Thus, our proposed framework can generate a massive dataset of that images are high-quality and ground truth-paired, which well addresses the needs of the annotated datasets in specific tasks.
Authors: Wonhyeok Choi, Mingyu Shin, Hyukzae Lee, Jaehoon Cho, Jaehyeon Park, Sunghoon Im
Abstract: Real-time processing is crucial in autonomous driving systems due to the imperative of instantaneous decision-making and rapid response. In real-world scenarios, autonomous vehicles are continuously tasked with interpreting their surroundings, analyzing intricate sensor data, and making decisions within split seconds to ensure safety through numerous computer vision tasks. In this paper, we present a new real-time multi-task network adept at three vital autonomous driving tasks: monocular 3D object detection, semantic segmentation, and dense depth estimation. To counter the challenge of negative transfer, which is the prevalent issue in multi-task learning, we introduce a task-adaptive attention generator. This generator is designed to automatically discern interrelations across the three tasks and arrange the task-sharing pattern, all while leveraging the efficiency of the hard-parameter sharing approach. To the best of our knowledge, the proposed model is pioneering in its capability to concurrently handle multiple tasks, notably 3D object detection, while maintaining real-time processing speeds. Our rigorously optimized network, when tested on the Cityscapes-3D datasets, consistently outperforms various baseline models. Moreover, an in-depth ablation study substantiates the efficacy of the methodologies integrated into our framework.
Authors: Yizheng Gong, Siyue Yu, Xiaoyang Wang, Jimin Xiao
Abstract: Most continual segmentation methods tackle the problem as a per-pixel classification task. However, such a paradigm is very challenging, and we find query-based segmenters with built-in objectness have inherent advantages compared with per-pixel ones, as objectness has strong transfer ability and forgetting resistance. Based on these findings, we propose CoMasTRe by disentangling continual segmentation into two stages: forgetting-resistant continual objectness learning and well-researched continual classification. CoMasTRe uses a two-stage segmenter learning class-agnostic mask proposals at the first stage and leaving recognition to the second stage. During continual learning, a simple but effective distillation is adopted to strengthen objectness. To further mitigate the forgetting of old classes, we design a multi-label class distillation strategy suited for segmentation. We assess the effectiveness of CoMasTRe on PASCAL VOC and ADE20K. Extensive experiments show that our method outperforms per-pixel and query-based methods on both datasets. Code will be available at https://github.com/jordangong/CoMasTRe.
Authors: Takahiro Shirakawa, Seiichi Uchida
Abstract: Layout-aware text-to-image generation is a task to generate multi-object images that reflect layout conditions in addition to text conditions. The current layout-aware text-to-image diffusion models still have several issues, including mismatches between the text and layout conditions and quality degradation of generated images. This paper proposes a novel layout-aware text-to-image diffusion model called NoiseCollage to tackle these issues. During the denoising process, NoiseCollage independently estimates noises for individual objects and then crops and merges them into a single noise. This operation helps avoid condition mismatches; in other words, it can put the right objects in the right places. Qualitative and quantitative evaluations show that NoiseCollage outperforms several state-of-the-art models. These successful results indicate that the crop-and-merge operation of noises is a reasonable strategy to control image generation. We also show that NoiseCollage can be integrated with ControlNet to use edges, sketches, and pose skeletons as additional conditions. Experimental results show that this integration boosts the layout accuracy of ControlNet. The code is available at https://github.com/univ-esuty/noisecollage.
Authors: Liang Peng, Junyuan Gao, Xinran Liu, Weihong Li, Shaohua Dong, Zhipeng Zhang, Heng Fan, Libo Zhang
Abstract: In this paper, we introduce a novel benchmark, dubbed VastTrack, towards facilitating the development of more general visual tracking via encompassing abundant classes and videos. VastTrack possesses several attractive properties: (1) Vast Object Category. In particular, it covers target objects from 2,115 classes, largely surpassing object categories of existing popular benchmarks (e.g., GOT-10k with 563 classes and LaSOT with 70 categories). With such vast object classes, we expect to learn more general object tracking. (2) Larger scale. Compared with current benchmarks, VastTrack offers 50,610 sequences with 4.2 million frames, which makes it to date the largest benchmark regarding the number of videos, and thus could benefit training even more powerful visual trackers in the deep learning era. (3) Rich Annotation. Besides conventional bounding box annotations, VastTrack also provides linguistic descriptions for the videos. The rich annotations of VastTrack enables development of both the vision-only and the vision-language tracking. To ensure precise annotation, all videos are manually labeled with multiple rounds of careful inspection and refinement. To understand performance of existing trackers and to provide baselines for future comparison, we extensively assess 25 representative trackers. The results, not surprisingly, show significant drops compared to those on current datasets due to lack of abundant categories and videos from diverse scenarios for training, and more efforts are required to improve general tracking. Our VastTrack and all the evaluation results will be made publicly available https://github.com/HengLan/VastTrack.
Authors: Lu Wen, Zhenghao Feng, Yun Hou, Peng Wang, Xi Wu, Jiliu Zhou, Yan Wang
Abstract: Semi-supervised learning is a sound measure to relieve the strict demand of abundant annotated datasets, especially for challenging multi-organ segmentation . However, most existing SSL methods predict pixels in a single image independently, ignoring the relations among images and categories. In this paper, we propose a two-stage Dual Contrastive Learning Network for semi-supervised MoS, which utilizes global and local contrastive learning to strengthen the relations among images and classes. Concretely, in Stage 1, we develop a similarity-guided global contrastive learning to explore the implicit continuity and similarity among images and learn global context. Then, in Stage 2, we present an organ-aware local contrastive learning to further attract the class representations. To ease the computation burden, we introduce a mask center computation algorithm to compress the category representations for local contrastive learning. Experiments conducted on the public 2017 ACDC dataset and an in-house RC-OARs dataset has demonstrated the superior performance of our method.
Authors: Quan Liu, Hongzi Zhu, Zhenxi Wang, Yunsong Zhou, Shan Chang, Minyi Guo
Abstract: Registration of point clouds collected from a pair of distant vehicles provides a comprehensive and accurate 3D view of the driving scenario, which is vital for driving safety related applications, yet existing literature suffers from the expensive pose label acquisition and the deficiency to generalize to new data distributions. In this paper, we propose EYOC, an unsupervised distant point cloud registration method that adapts to new point cloud distributions on the fly, requiring no global pose labels. The core idea of EYOC is to train a feature extractor in a progressive fashion, where in each round, the feature extractor, trained with near point cloud pairs, can label slightly farther point cloud pairs, enabling self-supervision on such far point cloud pairs. This process continues until the derived extractor can be used to register distant point clouds. Particularly, to enable high-fidelity correspondence label generation, we devise an effective spatial filtering scheme to select the most representative correspondences to register a point cloud pair, and then utilize the aligned point clouds to discover more correct correspondences. Experiments show that EYOC can achieve comparable performance with state-of-the-art supervised methods at a lower training cost. Moreover, it outwits supervised methods regarding generalization performance on new data distributions.
Authors: Minyang Hu, Hong Chang, Zong Guo, Bingpeng Ma, Shiguan Shan, Xilin Chen
Abstract: Few-shot learning (FSL) aims to learn novel tasks with very few labeled samples by leveraging experience from \emph{related} training tasks. In this paper, we try to understand FSL by delving into two key questions: (1) How to quantify the relationship between \emph{training} and \emph{novel} tasks? (2) How does the relationship affect the \emph{adaptation difficulty} on novel tasks for different models? To answer the two questions, we introduce Task Attribute Distance (TAD) built upon attributes as a metric to quantify the task relatedness. Unlike many existing metrics, TAD is model-agnostic, making it applicable to different FSL models. Then, we utilize TAD metric to establish a theoretical connection between task relatedness and task adaptation difficulty. By deriving the generalization error bound on a novel task, we discover how TAD measures the adaptation difficulty on novel tasks for FSL models. To validate our TAD metric and theoretical findings, we conduct experiments on three benchmarks. Our experimental results confirm that TAD metric effectively quantifies the task relatedness and reflects the adaptation difficulty on novel tasks for various FSL methods, even if some of them do not learn attributes explicitly or human-annotated attributes are not available. Finally, we present two applications of the proposed TAD metric: data augmentation and test-time intervention, which further verify its effectiveness and general applicability. The source code is available at https://github.com/hu-my/TaskAttributeDistance.
Authors: Peng Dai, Yang Zhang, Tao Liu, Zhen Fan, Tianyuan Du, Zhuo Su, Xiaozheng Zheng, Zeming Li
Abstract: It is especially challenging to achieve real-time human motion tracking on a standalone VR Head-Mounted Display (HMD) such as Meta Quest and PICO. In this paper, we propose HMD-Poser, the first unified approach to recover full-body motions using scalable sparse observations from HMD and body-worn IMUs. In particular, it can support a variety of input scenarios, such as HMD, HMD+2IMUs, HMD+3IMUs, etc. The scalability of inputs may accommodate users' choices for both high tracking accuracy and easy-to-wear. A lightweight temporal-spatial feature learning network is proposed in HMD-Poser to guarantee that the model runs in real-time on HMDs. Furthermore, HMD-Poser presents online body shape estimation to improve the position accuracy of body joints. Extensive experimental results on the challenging AMASS dataset show that HMD-Poser achieves new state-of-the-art results in both accuracy and real-time performance. We also build a new free-dancing motion dataset to evaluate HMD-Poser's on-device performance and investigate the performance gap between synthetic data and real-captured sensor data. Finally, we demonstrate our HMD-Poser with a real-time Avatar-driving application on a commercial HMD. Our code and free-dancing motion dataset are available https://pico-ai-team.github.io/hmd-poser
Authors: Zi-Ting Chou, Sheng-Yu Huang, I-Jieh Liu, Yu-Chiang Frank Wang
Abstract: Utilizing multi-view inputs to synthesize novel-view images, Neural Radiance Fields (NeRF) have emerged as a popular research topic in 3D vision. In this work, we introduce a Generalizable Semantic Neural Radiance Field (GSNeRF), which uniquely takes image semantics into the synthesis process so that both novel view images and the associated semantic maps can be produced for unseen scenes. Our GSNeRF is composed of two stages: Semantic Geo-Reasoning and Depth-Guided Visual rendering. The former is able to observe multi-view image inputs to extract semantic and geometry features from a scene. Guided by the resulting image geometry information, the latter performs both image and semantic rendering with improved performances. Our experiments not only confirm that GSNeRF performs favorably against prior works on both novel-view image and semantic segmentation synthesis but the effectiveness of our sampling strategy for visual rendering is further verified.
Authors: Muhammad Kashif Ali, Eun Woo Im, Dongjin Kim, Tae Hyun Kim
Abstract: Video stabilization is a longstanding computer vision problem, particularly pixel-level synthesis solutions for video stabilization which synthesize full frames add to the complexity of this task. These techniques aim to stabilize videos by synthesizing full frames while enhancing the stability of the considered video. This intensifies the complexity of the task due to the distinct mix of unique motion profiles and visual content present in each video sequence, making robust generalization with fixed parameters difficult. In our study, we introduce a novel approach to enhance the performance of pixel-level synthesis solutions for video stabilization by adapting these models to individual input video sequences. The proposed adaptation exploits low-level visual cues accessible during test-time to improve both the stability and quality of resulting videos. We highlight the efficacy of our methodology of "test-time adaptation" through simple fine-tuning of one of these models, followed by significant stability gain via the integration of meta-learning techniques. Notably, significant improvement is achieved with only a single adaptation step. The versatility of the proposed algorithm is demonstrated by consistently improving the performance of various pixel-level synthesis models for video stabilization in real-world scenarios.
Authors: Xavier Bou, Thibaud Ehret, Rafael Grompone von Gioi, Jeremy Anger
Abstract: Identifying flood affected areas in remote sensing data is a critical problem in earth observation to analyze flood impact and drive responses. While a number of methods have been proposed in the literature, there are two main limitations in available flood detection datasets: (1) a lack of region variability is commonly observed and/or (2) they require to distinguish permanent water bodies from flooded areas from a single image, which becomes an ill-posed setup. Consequently, we extend the globally diverse MMFlood dataset to multi-date by providing one year of Sentinel-1 observations around each flood event. To our surprise, we notice that the definition of flooded pixels in MMFlood is inconsistent when observing the entire image sequence. Hence, we re-frame the flood detection task as a temporal anomaly detection problem, where anomalous water bodies are segmented from a Sentinel-1 temporal sequence. From this definition, we provide a simple method inspired by the popular video change detector ViBe, results of which quantitatively align with the SAR image time series, providing a reasonable baseline for future works.
Authors: Kalibinuer Tiliwalidi
Abstract: Currently, infrared imaging technology enjoys widespread usage, with infrared object detection technology experiencing a surge in prominence. While previous studies have delved into physical attacks on infrared object detectors, the implementation of these techniques remains complex. For instance, some approaches entail the use of bulb boards or infrared QR suits as perturbations to execute attacks, which entail costly optimization and cumbersome deployment processes. Other methodologies involve the utilization of irregular aerogel as physical perturbations for infrared attacks, albeit at the expense of optimization expenses and perceptibility issues. In this study, we propose a novel infrared physical attack termed Adversarial Infrared Geometry (\textbf{AdvIG}), which facilitates efficient black-box query attacks by modeling diverse geometric shapes (lines, triangles, ellipses) and optimizing their physical parameters using Particle Swarm Optimization (PSO). Extensive experiments are conducted to evaluate the effectiveness, stealthiness, and robustness of AdvIG. In digital attack experiments, line, triangle, and ellipse patterns achieve attack success rates of 93.1\%, 86.8\%, and 100.0\%, respectively, with average query times of 71.7, 113.1, and 2.57, respectively, thereby confirming the efficiency of AdvIG. Physical attack experiments are conducted to assess the attack success rate of AdvIG at different distances. On average, the line, triangle, and ellipse achieve attack success rates of 61.1\%, 61.2\%, and 96.2\%, respectively. Further experiments are conducted to comprehensively analyze AdvIG, including ablation experiments, transfer attack experiments, and adversarial defense mechanisms. Given the superior performance of our method as a simple and efficient black-box adversarial attack in both digital and physical environments, we advocate for widespread attention to AdvIG.
Authors: Yufan Chen, Ching Ting Leung, Yong Huang, Jianwei Sun, Hao Chen, Hanyu Gao
Abstract: In the field of chemical structure recognition, the task of converting molecular images into graph structures and SMILES string stands as a significant challenge, primarily due to the varied drawing styles and conventions prevalent in chemical literature. To bridge this gap, we proposed MolNexTR, a novel image-to-graph deep learning model that collaborates to fuse the strengths of ConvNext, a powerful Convolutional Neural Network variant, and Vision-TRansformer. This integration facilitates a more nuanced extraction of both local and global features from molecular images. MolNexTR can predict atoms and bonds simultaneously and understand their layout rules. It also excels at flexibly integrating symbolic chemistry principles to discern chirality and decipher abbreviated structures. We further incorporate a series of advanced algorithms, including improved data augmentation module, image contamination module, and a post-processing module to get the final SMILES output. These modules synergistically enhance the model's robustness against the diverse styles of molecular imagery found in real literature. In our test sets, MolNexTR has demonstrated superior performance, achieving an accuracy rate of 81-97%, marking a significant advancement in the domain of molecular structure recognition. Scientific contribution: MolNexTR is a novel image-to-graph model that incorporates a unique dual-stream encoder to extract complex molecular image features, and combines chemical rules to predict atoms and bonds while understanding atom and bond layout rules. In addition, it employs a series of novel augmentation algorithms to significantly enhance the robustness and performance of the model.
Authors: Jingru Zhu, Ya Guo, Geng Sun, Liang Hong, Jie Chen
Abstract: Semantic segmentation of high-resolution remote sensing imagery (HRSI) suffers from the domain shift, resulting in poor performance of the model in another unseen domain. Unsupervised domain adaptive (UDA) semantic segmentation aims to adapt the semantic segmentation model trained on the labeled source domain to an unlabeled target domain. However, the existing UDA semantic segmentation models tend to align pixels or features based on statistical information related to labels in source and target domain data, and make predictions accordingly, which leads to uncertainty and fragility of prediction results. In this paper, we propose a causal prototype-inspired contrast adaptation (CPCA) method to explore the invariant causal mechanisms between different HRSIs domains and their semantic labels. It firstly disentangles causal features and bias features from the source and target domain images through a causal feature disentanglement module. Then, a causal prototypical contrast module is used to learn domain invariant causal features. To further de-correlate causal and bias features, a causal intervention module is introduced to intervene on the bias features to generate counterfactual unbiased samples. By forcing the causal features to meet the principles of separability, invariance and intervention, CPCA can simulate the causal factors of source and target domains, and make decisions on the target domain based on the causal features, which can observe improved generalization ability. Extensive experiments under three cross-domain tasks indicate that CPCA is remarkably superior to the state-of-the-art methods.
Authors: Yajie Liu, Pu Ge, Qingjie Liu, Di Huang
Abstract: Recently, learning open-vocabulary semantic segmentation from text supervision has achieved promising downstream performance. Nevertheless, current approaches encounter an alignment granularity gap owing to the absence of dense annotations, wherein they learn coarse image/region-text alignment during training yet perform group/pixel-level predictions at inference. Such discrepancy leads to suboptimal learning efficiency and inferior zero-shot segmentation results. In this paper, we introduce a Multi-Grained Cross-modal Alignment (MGCA) framework, which explicitly learns pixel-level alignment along with object- and region-level alignment to bridge the granularity gap without any dense annotations. Specifically, MGCA ingeniously constructs pseudo multi-granular semantic correspondences upon image-text pairs and collaborates with hard sampling strategies to facilitate fine-grained cross-modal contrastive learning. Further, we point out the defects of existing group and pixel prediction units in downstream segmentation and develop an adaptive semantic unit which effectively mitigates their dilemmas including under- and over-segmentation. Training solely on CC3M, our method achieves significant advancements over state-of-the-art methods, demonstrating its effectiveness and efficiency.
Authors: Zequn Zeng, Yan Xie, Hao Zhang, Chiyu Chen, Zhengjue Wang, Bo Chen
Abstract: Zero-shot image captioning (IC) without well-paired image-text data can be divided into two categories, training-free and text-only-training. Generally, these two types of methods realize zero-shot IC by integrating pretrained vision-language models like CLIP for image-text similarity evaluation and a pre-trained language model (LM) for caption generation. The main difference between them is whether using a textual corpus to train the LM. Though achieving attractive performance w.r.t. some metrics, existing methods often exhibit some common drawbacks. Training-free methods tend to produce hallucinations, while text-only-training often lose generalization capability. To move forward, in this paper, we propose a novel Memory-Augmented zero-shot image Captioning framework (MeaCap). Specifically, equipped with a textual memory, we introduce a retrieve-then-filter module to get key concepts that are highly related to the image. By deploying our proposed memory-augmented visual-related fusion score in a keywords-to-sentence LM, MeaCap can generate concept-centered captions that keep high consistency with the image with fewer hallucinations and more world-knowledge. The framework of MeaCap achieves the state-of-the-art performance on a series of zero-shot IC settings. Our code is available at https://github.com/joeyz0z/MeaCap.
Authors: Emanuele Vivoli, Joan Lafuente Baeza, Ernest Valveny Llobet, Dimosthenis Karatzas
Abstract: This work explores a closure task in comics, a medium where visual and textual elements are intricately intertwined. Specifically, Text-cloze refers to the task of selecting the correct text to use in a comic panel, given its neighboring panels. Traditional methods based on recurrent neural networks have struggled with this task due to limited OCR accuracy and inherent model limitations. We introduce a novel Multimodal Large Language Model (Multimodal-LLM) architecture, specifically designed for Text-cloze, achieving a 10% improvement over existing state-of-the-art models in both its easy and hard variants. Central to our approach is a Domain-Adapted ResNet-50 based visual encoder, fine-tuned to the comics domain in a self-supervised manner using SimCLR. This encoder delivers comparable results to more complex models with just one-fifth of the parameters. Additionally, we release new OCR annotations for this dataset, enhancing model input quality and resulting in another 1% improvement. Finally, we extend the task to a generative format, establishing new baselines and expanding the research possibilities in the field of comics analysis.
Authors: Gyusam Chang, Wonseok Roh, Sujin Jang, Dongwook Lee, Daehyun Ji, Gyeongrok Oh, Jinsun Park, Jinkyu Kim, Sangpil Kim
Abstract: Recent LiDAR-based 3D Object Detection (3DOD) methods show promising results, but they often do not generalize well to target domains outside the source (or training) data distribution. To reduce such domain gaps and thus to make 3DOD models more generalizable, we introduce a novel unsupervised domain adaptation (UDA) method, called CMDA, which (i) leverages visual semantic cues from an image modality (i.e., camera images) as an effective semantic bridge to close the domain gap in the cross-modal Bird's Eye View (BEV) representations. Further, (ii) we also introduce a self-training-based learning strategy, wherein a model is adversarially trained to generate domain-invariant features, which disrupt the discrimination of whether a feature instance comes from a source or an unseen target domain. Overall, our CMDA framework guides the 3DOD model to generate highly informative and domain-adaptive features for novel data distributions. In our extensive experiments with large-scale benchmarks, such as nuScenes, Waymo, and KITTI, those mentioned above provide significant performance gains for UDA tasks, achieving state-of-the-art performance.
Authors: John Day, Tushar Arora, Jirui Liu, Li Erran Li, Ming Bo Cai
Abstract: As part of human core knowledge, the representation of objects is the building block of mental representation that supports high-level concepts and symbolic reasoning. While humans develop the ability of perceiving objects situated in 3D environments without supervision, models that learn the same set of abilities with similar constraints faced by human infants are lacking. Towards this end, we developed a novel network architecture that simultaneously learns to 1) segment objects from discrete images, 2) infer their 3D locations, and 3) perceive depth, all while using only information directly available to the brain as training data, namely: sequences of images and self-motion. The core idea is treating objects as latent causes of visual input which the brain uses to make efficient predictions of future scenes. This results in object representations being learned as an essential byproduct of learning to predict.
Authors: Naifu Xue, Qi Mao, Zijian Wang, Yuan Zhang, Siwei Ma
Abstract: Recent progress in generative compression technology has significantly improved the perceptual quality of compressed data. However, these advancements primarily focus on producing high-frequency details, often overlooking the ability of generative models to capture the prior distribution of image content, thus impeding further bitrate reduction in extreme compression scenarios (<0.05 bpp). Motivated by the capabilities of predictive language models for lossless compression, this paper introduces a novel Unified Image Generation-Compression (UIGC) paradigm, merging the processes of generation and compression. A key feature of the UIGC framework is the adoption of vector-quantized (VQ) image models for tokenization, alongside a multi-stage transformer designed to exploit spatial contextual information for modeling the prior distribution. As such, the dual-purpose framework effectively utilizes the learned prior for entropy estimation and assists in the regeneration of lost tokens. Extensive experiments demonstrate the superiority of the proposed UIGC framework over existing codecs in perceptual quality and human perception, particularly in ultra-low bitrate scenarios (<=0.03 bpp), pioneering a new direction in generative compression.
Authors: Zhaoran Zhao, Peng Lu, Xujun Peng, Wenhao Guo
Abstract: In the domain of image layout representation learning, the critical process of translating image layouts into succinct vector forms is increasingly significant across diverse applications, such as image retrieval, manipulation, and generation. Most approaches in this area heavily rely on costly labeled datasets and notably lack in adapting their modeling and learning methods to the specific nuances of photographic image layouts. This shortfall makes the learning process for photographic image layouts suboptimal. In our research, we directly address these challenges. We innovate by defining basic layout primitives that encapsulate various levels of layout information and by mapping these, along with their interconnections, onto a heterogeneous graph structure. This graph is meticulously engineered to capture the intricate layout information within the pixel domain explicitly. Advancing further, we introduce novel pretext tasks coupled with customized loss functions, strategically designed for effective self-supervised learning of these layout graphs. Building on this foundation, we develop an autoencoder-based network architecture skilled in compressing these heterogeneous layout graphs into precise, dimensionally-reduced layout representations. Additionally, we introduce the LODB dataset, which features a broader range of layout categories and richer semantics, serving as a comprehensive benchmark for evaluating the effectiveness of layout representation learning methods. Our extensive experimentation on this dataset demonstrates the superior performance of our approach in the realm of photographic image layout representation learning.
Authors: Wei Zhang, Miaoxin Cai, Tong Zhang, Guoqiang Lei, Yin Zhuang, Xuerui Mao
Abstract: Ship detection needs to identify ship locations from remote sensing (RS) scenes. However, due to different imaging payloads, various appearances of ships, and complicated background interference from the bird's eye view, it is difficult to set up a unified paradigm for achieving multi-source ship detection. Therefore, in this article, considering that the large language models (LLMs) emerge the powerful generalization ability, a novel unified visual-language model called Popeye is proposed for multi-source ship detection from RS imagery. First, to bridge the interpretation gap between multi-source images for ship detection, a novel image-instruction-answer way is designed to integrate the various ship detection ways (e.g., horizontal bounding box (HBB), oriented bounding box (OBB)) into a unified labeling paradigm. Then, in view of this, a cross-modal image interpretation method is developed for the proposed Popeye to enhance interactive comprehension ability between visual and language content, which can be easily migrated into any multi-source ship detection task. Subsequently, owing to objective domain differences, a knowledge adaption mechanism is designed to adapt the pre-trained visual-language knowledge from the nature scene into the RS domain for multi-source ship detection. In addition, the segment anything model (SAM) is also seamlessly integrated into the proposed Popeye to achieve pixel-level ship segmentation without additional training costs. Finally, extensive experiments are conducted on the newly constructed instruction dataset named MMShip, and the results indicate that the proposed Popeye outperforms current specialist, open-vocabulary, and other visual-language models for zero-shot multi-source ship detection.
Authors: Jeremias Gerner, Klaus Bogenberger, Stefanie Schmidtner
Abstract: Floating Car Observers (FCOs) are an innovative method to collect traffic data by deploying sensor-equipped vehicles to detect and locate other vehicles. We demonstrate that even a small penetration rate of FCOs can identify a significant amount of vehicles at a given intersection. This is achieved through the emulation of detection within a microscopic traffic simulation. Additionally, leveraging data from previous moments can enhance the detection of vehicles in the current frame. Our findings indicate that, with a 20-second observation window, it is possible to recover up to 20\% of vehicles that are not visible by FCOs in the current timestep. To exploit this, we developed a data-driven strategy, utilizing sequences of Bird's Eye View (BEV) representations of detected vehicles and deep learning models. This approach aims to bring currently undetected vehicles into view in the present moment, enhancing the currently detected vehicles. Results of different spatiotemporal architectures show that up to 41\% of the vehicles can be recovered into the current timestep at their current position. This enhancement enriches the information initially available by the FCO, allowing an improved estimation of traffic states and metrics (e.g. density and queue length) for improved implementation of traffic management strategies.
Authors: Erik Brorsson, Knut {\AA}kesson, Lennart Svensson, Kristofer Bengtsson
Abstract: We consider unsupervised domain adaptation (UDA) for semantic segmentation in which the model is trained on a labeled source dataset and adapted to an unlabeled target dataset. Unfortunately, current self-training methods are susceptible to misclassified pseudo-labels resulting from erroneous predictions. Since certain classes are typically associated with less reliable predictions in UDA, reducing the impact of such pseudo-labels without skewing the training towards some classes is notoriously difficult. To this end, we propose an extensive cut-and-paste strategy (ECAP) to leverage reliable pseudo-labels through data augmentation. Specifically, ECAP maintains a memory bank of pseudo-labeled target samples throughout training and cut-and-pastes the most confident ones onto the current training batch. We implement ECAP on top of the recent method MIC and boost its performance on two synthetic-to-real domain adaptation benchmarks. Notably, MIC+ECAP reaches an unprecedented performance of 69.1 mIoU on the Synthia->Cityscapes benchmark. Our code is available at https://github.com/ErikBrorsson/ECAP.
Authors: Deepanway Ghosal, Vernon Toh Yan Han, Chia Yew Ken, Soujanya Poria
Abstract: This paper introduces the novel task of multimodal puzzle solving, framed within the context of visual question-answering. We present a new dataset, AlgoPuzzleVQA designed to challenge and evaluate the capabilities of multimodal language models in solving algorithmic puzzles that necessitate both visual understanding, language understanding, and complex algorithmic reasoning. We create the puzzles to encompass a diverse array of mathematical and algorithmic topics such as boolean logic, combinatorics, graph theory, optimization, search, etc., aiming to evaluate the gap between visual data interpretation and algorithmic problem-solving skills. The dataset is generated automatically from code authored by humans. All our puzzles have exact solutions that can be found from the algorithm without tedious human calculations. It ensures that our dataset can be scaled up arbitrarily in terms of reasoning complexity and dataset size. Our investigation reveals that large language models (LLMs) such as GPT4V and Gemini exhibit limited performance in puzzle-solving tasks. We find that their performance is near random in a multi-choice question-answering setup for a significant number of puzzles. The findings emphasize the challenges of integrating visual, language, and algorithmic knowledge for solving complex reasoning problems.
Authors: Meryem Amaouche, Ouassim Karrakchou, Mounir Ghogho, Anouar El Ghazzaly, Mohamed Alami, Ahmed Ameur
Abstract: Bladder cancer ranks within the top 10 most diagnosed cancers worldwide and is among the most expensive cancers to treat due to the high recurrence rates which require lifetime follow-ups. The primary tool for diagnosis is cystoscopy, which heavily relies on doctors' expertise and interpretation. Therefore, annually, numerous cases are either undiagnosed or misdiagnosed and treated as urinary infections. To address this, we suggest a deep learning approach for bladder cancer detection and segmentation which combines CNNs with a lightweight positional-encoding-free transformer and dual attention gates that fuse self and spatial attention for feature enhancement. The architecture suggested in this paper is efficient making it suitable for medical scenarios that require real time inference. Experiments have proven that this model addresses the critical need for a balance between computational efficiency and diagnostic accuracy in cystoscopic imaging as despite its small size it rivals large models in performance.
Authors: Brian B. Moser, Federico Raue, Sebastian Palacio, Stanislav Frolov, Andreas Dengel
Abstract: The efficacy of machine learning has traditionally relied on the availability of increasingly larger datasets. However, large datasets pose storage challenges and contain non-influential samples, which could be ignored during training without impacting the final accuracy of the model. In response to these limitations, the concept of distilling the information on a dataset into a condensed set of (synthetic) samples, namely a distilled dataset, emerged. One crucial aspect is the selected architecture (usually ConvNet) for linking the original and synthetic datasets. However, the final accuracy is lower if the employed model architecture differs from the model used during distillation. Another challenge is the generation of high-resolution images, e.g., 128x128 and higher. In this paper, we propose Latent Dataset Distillation with Diffusion Models (LD3M) that combine diffusion in latent space with dataset distillation to tackle both challenges. LD3M incorporates a novel diffusion process tailored for dataset distillation, which improves the gradient norms for learning synthetic images. By adjusting the number of diffusion steps, LD3M also offers a straightforward way of controlling the trade-off between speed and accuracy. We evaluate our approach in several ImageNet subsets and for high-resolution images (128x128 and 256x256). As a result, LD3M consistently outperforms state-of-the-art distillation techniques by up to 4.8 p.p. and 4.2 p.p. for 1 and 10 images per class, respectively.
Authors: Jianfei Liu, Christopher Parnell, Ronald M. Summers
Abstract: Accurate training labels are a key component for multi-class medical image segmentation. Their annotation is costly and time-consuming because it requires domain expertise. This work aims to develop a dual-branch network and automatically improve training labels for multi-class image segmentation. Transfer learning is used to train the network and improve inaccurate weak labels sequentially. The dual-branch network is first trained by weak labels alone to initialize model parameters. After the network is stabilized, the shared encoder is frozen, and strong and weak decoders are fine-tuned by strong and weak labels together. The accuracy of weak labels is iteratively improved in the fine-tuning process. The proposed method was applied to a three-class segmentation of muscle, subcutaneous and visceral adipose tissue on abdominal CT scans. Validation results on 11 patients showed that the accuracy of training labels was statistically significantly improved, with the Dice similarity coefficient of muscle, subcutaneous and visceral adipose tissue increased from 74.2% to 91.5%, 91.2% to 95.6%, and 77.6% to 88.5%, respectively (p<0.05). In comparison with our earlier method, the label accuracy was also significantly improved (p<0.05). These experimental results suggested that the combination of the dual-branch network and transfer learning is an efficient means to improve training labels for multi-class segmentation.
Authors: Tianshu Huang, John Miller, Akarsh Prabhakara, Tao Jin, Tarana Laroia, Zico Kolter, Anthony Rowe
Abstract: Simulation is an invaluable tool for radio-frequency system designers that enables rapid prototyping of various algorithms for imaging, target detection, classification, and tracking. However, simulating realistic radar scans is a challenging task that requires an accurate model of the scene, radio frequency material properties, and a corresponding radar synthesis function. Rather than specifying these models explicitly, we propose DART - Doppler Aided Radar Tomography, a Neural Radiance Field-inspired method which uses radar-specific physics to create a reflectance and transmittance-based rendering pipeline for range-Doppler images. We then evaluate DART by constructing a custom data collection platform and collecting a novel radar dataset together with accurate position and instantaneous velocity measurements from lidar-based localization. In comparison to state-of-the-art baselines, DART synthesizes superior radar range-Doppler images from novel views across all datasets and additionally can be used to generate high quality tomographic images.
Authors: Chang Liu, Fuxin Fan, Annette Schwarz, Andreas Maier
Abstract: Multi-organ segmentation in medical images is a widely researched task and can save much manual efforts of clinicians in daily routines. Automating the organ segmentation process using deep learning (DL) is a promising solution and state-of-the-art segmentation models are achieving promising accuracy. In this work, We proposed a novel data augmentation strategy for increasing the generalizibility of multi-organ segmentation datasets, namely AnatoMix. By object-level matching and manipulation, our method is able to generate new images with correct anatomy, i.e. organ segmentation mask, exponentially increasing the size of the segmentation dataset. Initial experiments have been done to investigate the segmentation performance influenced by our method on a public CT dataset. Our augmentation method can lead to mean dice of 76.1, compared with 74.8 of the baseline method.
Authors: Ganesh Babu, Aoife Gowen, Michael Fop, Isobel Claire Gormley
Abstract: The use of hyperspectral imaging to investigate food samples has grown due to the improved performance and lower cost of spectroscopy instrumentation. Food engineers use hyperspectral images to classify the type and quality of a food sample, typically using classification methods. In order to train these methods, every pixel in each training image needs to be labelled. Typically, computationally cheap threshold-based approaches are used to label the pixels, and classification methods are trained based on those labels. However, threshold-based approaches are subjective and cannot be generalized across hyperspectral images taken in different conditions and of different foods. Here a consensus-constrained parsimonious Gaussian mixture model (ccPGMM) is proposed to label pixels in hyperspectral images using a model-based clustering approach. The ccPGMM utilizes available information on the labels of a small number of pixels and the relationship between those pixels and neighbouring pixels as constraints when clustering the rest of the pixels in the image. A latent variable model is used to represent the high-dimensional data in terms of a small number of underlying latent factors. To ensure computational feasibility, a consensus clustering approach is employed, where the data are divided into multiple randomly selected subsets of variables and constrained clustering is applied to each data subset; the clustering results are then consolidated across all data subsets to provide a consensus clustering solution. The ccPGMM approach is applied to simulated datasets and real hyperspectral images of three types of puffed cereal, corn, rice, and wheat. Improved clustering performance and computational efficiency are demonstrated when compared to other current state-of-the-art approaches.
Authors: Chen Yuhua
Abstract: This article investigates deep learning methodologies for single-modality clinical data analysis, as a crucial precursor to multi-modal medical research. Building on Guo JingYuan's work, the study refines clinical data processing through Compact Convolutional Transformer (CCT), Patch Up, and the innovative CamCenterLoss technique, establishing a foundation for future multimodal investigations. The proposed methodology demonstrates improved prediction accuracy and at tentiveness to critically ill patients compared to Guo JingYuan's ResNet and StageNet approaches. Novelty that using image-pretrained vision transformer backbone to perform transfer learning time-series clinical data.The study highlights the potential of CCT, Patch Up, and novel CamCenterLoss in processing single modality clinical data within deep learning frameworks, paving the way for future multimodal medical research and promoting precision and personalized healthcare
Authors: Yingrui Ji, Yao Zhu, Zhigang Li, Jiansheng Chen, Yunlong Kong, Jingbo Chen
Abstract: In the dynamic realms of machine learning and deep learning, the robustness and reliability of models are paramount, especially in critical real-world applications. A fundamental challenge in this sphere is managing Out-of-Distribution (OOD) samples, significantly increasing the risks of model misclassification and uncertainty. Our work addresses this challenge by enhancing the detection and management of OOD samples in neural networks. We introduce OOD-R (Out-of-Distribution-Rectified), a meticulously curated collection of open-source datasets with enhanced noise reduction properties. In-Distribution (ID) noise in existing OOD datasets can lead to inaccurate evaluation of detection algorithms. Recognizing this, OOD-R incorporates noise filtering technologies to refine the datasets, ensuring a more accurate and reliable evaluation of OOD detection algorithms. This approach not only improves the overall quality of data but also aids in better distinguishing between OOD and ID samples, resulting in up to a 2.5\% improvement in model accuracy and a minimum 3.2\% reduction in false positives. Furthermore, we present ActFun, an innovative method that fine-tunes the model's response to diverse inputs, thereby improving the stability of feature extraction and minimizing specificity issues. ActFun addresses the common problem of model overconfidence in OOD detection by strategically reducing the influence of hidden units, which enhances the model's capability to estimate OOD uncertainty more accurately. Implementing ActFun in the OOD-R dataset has led to significant performance enhancements, including an 18.42\% increase in AUROC of the GradNorm method and a 16.93\% decrease in FPR95 of the Energy method. Overall, our research not only advances the methodologies in OOD detection but also emphasizes the importance of dataset integrity for accurate algorithm evaluation.
Authors: Rina Su, Yu Guo, Caiying Wu, Qiyu Jin, Tieyong Zeng
Abstract: The main objective of the Multiple Kernel k-Means (MKKM) algorithm is to extract non-linear information and achieve optimal clustering by optimizing base kernel matrices. Current methods enhance information diversity and reduce redundancy by exploiting interdependencies among multiple kernels based on correlations or dissimilarities. Nevertheless, relying solely on a single metric, such as correlation or dissimilarity, to define kernel relationships introduces bias and incomplete characterization. Consequently, this limitation hinders efficient information extraction, ultimately compromising clustering performance. To tackle this challenge, we introduce a novel method that systematically integrates both kernel correlation and dissimilarity. Our approach comprehensively captures kernel relationships, facilitating more efficient classification information extraction and improving clustering performance. By emphasizing the coherence between kernel correlation and dissimilarity, our method offers a more objective and transparent strategy for extracting non-linear information and significantly improving clustering precision, supported by theoretical rationale. We assess the performance of our algorithm on 13 challenging benchmark datasets, demonstrating its superiority over contemporary state-of-the-art MKKM techniques.
Authors: Ali Ayub, Chrystopher Nehaniv, Kerstin Dautenhahn
Abstract: For robots to perform assistive tasks in unstructured home environments, they must learn and reason on the semantic knowledge of the environments. Despite a resurgence in the development of semantic reasoning architectures, these methods assume that all the training data is available a priori. However, each user's environment is unique and can continue to change over time, which makes these methods unsuitable for personalized home service robots. Although research in continual learning develops methods that can learn and adapt over time, most of these methods are tested in the narrow context of object classification on static image datasets. In this paper, we combine ideas from continual learning, semantic reasoning, and interactive machine learning literature and develop a novel interactive continual learning architecture for continual learning of semantic knowledge in a home environment through human-robot interaction. The architecture builds on core cognitive principles of learning and memory for efficient and real-time learning of new knowledge from humans. We integrate our architecture with a physical mobile manipulator robot and perform extensive system evaluations in a laboratory environment over two months. Our results demonstrate the effectiveness of our architecture to allow a physical robot to continually adapt to the changes in the environment from limited data provided by the users (experimenters), and use the learned knowledge to perform object fetching tasks.
Authors: Weihao Jiang, Guodong Liu, Di He, Kun He
Abstract: Few-shot learning, a challenging task in machine learning, aims to learn a classifier adaptable to recognize new, unseen classes with limited labeled examples. Meta-learning has emerged as a prominent framework for few-shot learning. Its training framework is originally a task-level learning method, such as Model-Agnostic Meta-Learning (MAML) and Prototypical Networks. And a recently proposed training paradigm called Meta-Baseline, which consists of sequential pre-training and meta-training stages, gains state-of-the-art performance. However, as a non-end-to-end training method, indicating the meta-training stage can only begin after the completion of pre-training, Meta-Baseline suffers from higher training cost and suboptimal performance due to the inherent conflicts of the two training stages. To address these limitations, we propose an end-to-end training paradigm consisting of two alternative loops. In the outer loop, we calculate cross entropy loss on the entire training set while updating only the final linear layer. In the inner loop, we employ the original meta-learning training mode to calculate the loss and incorporate gradients from the outer loss to guide the parameter updates. This training paradigm not only converges quickly but also outperforms existing baselines, indicating that information from the overall training set and the meta-learning training paradigm could mutually reinforce one another. Moreover, being model-agnostic, our framework achieves significant performance gains, surpassing the baseline systems by approximate 1%.
Authors: Xinwei Ou, Ce Zhu, Xiaolin Huang, Yipeng Liu
Abstract: Second-order methods can converge much faster than first-order methods by incorporating second-order derivates or statistics, but they are far less prevalent in deep learning due to their computational inefficiency. To handle this, many of the existing solutions focus on reducing the size of the matrix to be inverted. However, it is still needed to perform the inverse operator in each iteration. In this paper, we present a fast natural gradient descent (FNGD) method, which only requires computing the inverse during the first epoch. Firstly, we reformulate the gradient preconditioning formula in the natural gradient descent (NGD) as a weighted sum of per-sample gradients using the Sherman-Morrison-Woodbury formula. Building upon this, to avoid the iterative inverse operation involved in computing coefficients, the weighted coefficients are shared across epochs without affecting the empirical performance. FNGD approximates the NGD as a fixed-coefficient weighted sum, akin to the average sum in first-order methods. Consequently, the computational complexity of FNGD can approach that of first-order methods. To demonstrate the efficiency of the proposed FNGD, we perform empirical evaluations on image classification and machine translation tasks. For training ResNet-18 on the CIFAR-100 dataset, FNGD can achieve a speedup of 2.05$\times$ compared with KFAC. For training Transformer on Multi30K, FNGD outperforms AdamW by 24 BLEU score while requiring almost the same training time.
Authors: Yu Guo, Axel Davy, Gabriele Facciolo, Jean-Michel Morel, Qiyu Jin
Abstract: With the widespread application of convolutional neural networks (CNNs), the traditional model based denoising algorithms are now outperformed. However, CNNs face two problems. First, they are computationally demanding, which makes their deployment especially difficult for mobile terminals. Second, experimental evidence shows that CNNs often over-smooth regular textures present in images, in contrast to traditional non-local models. In this letter, we propose a solution to both issues by combining a nonlocal algorithm with a lightweight residual CNN. This solution gives full latitude to the advantages of both models. We apply this framework to two GPU implementations of classic nonlocal algorithms (NLM and BM3D) and observe a substantial gain in both cases, performing better than the state-of-the-art with low computational requirements. Our solution is between 10 and 20 times faster than CNNs with equivalent performance and attains higher PSNR. In addition the final method shows a notable gain on images containing complex textures like the ones of the MIT Moire dataset.
Authors: Thomas Pinetz, Erich Kobler, Robert Haase, Julian A. Luetkens, Mathias Meetschen, Johannes Haubold, Cornelius Deuschl, Alexander Radbruch, Katerina Deike, Alexander Effland
Abstract: Recently, deep learning (DL)-based methods have been proposed for the computational reduction of gadolinium-based contrast agents (GBCAs) to mitigate adverse side effects while preserving diagnostic value. Currently, the two main challenges for these approaches are the accurate prediction of contrast enhancement and the synthesis of realistic images. In this work, we address both challenges by utilizing the contrast signal encoded in the subtraction images of pre-contrast and post-contrast image pairs. To avoid the synthesis of any noise or artifacts and solely focus on contrast signal extraction and enhancement from low-dose subtraction images, we train our DL model using noise-free standard-dose subtraction images as targets. As a result, our model predicts the contrast enhancement signal only; thereby enabling synthesization of images beyond the standard dose. Furthermore, we adapt the embedding idea of recent diffusion-based models to condition our model on physical parameters affecting the contrast enhancement behavior. We demonstrate the effectiveness of our approach on synthetic and real datasets using various scanners, field strengths, and contrast agents.
Authors: Tim Selig, Thomas M\"arz, Martin Storath, Andreas Weinmann
Abstract: Computed Tomography (CT) is a widely used medical imaging modality, and as it is based on ionizing radiation, it is desirable to minimize the radiation dose. However, a reduced radiation dose comes with reduced image quality, and reconstruction from low-dose CT (LDCT) data is still a challenging task which is subject to research. According to the LoDoPaB-CT benchmark, a benchmark for LDCT reconstruction, many state-of-the-art methods use pipelines involving UNet-type architectures. Specifically the top ranking method, ItNet, employs a three-stage process involving filtered backprojection (FBP), a UNet trained on CT data, and an iterative refinement step. In this paper, we propose a less complex two-stage method. The first stage also employs FBP, while the novelty lies in the training strategy for the second stage, characterized as the CT image enhancement stage. The crucial point of our approach is that the neural network is pretrained on a distinctly different pretraining task with non-CT data, namely Gaussian noise removal on a variety of natural grayscale images (photographs). We then fine-tune this network for the downstream task of CT image enhancement using pairs of LDCT images and corresponding normal-dose CT images (NDCT). Despite being notably simpler than the state-of-the-art, as the pretraining did not depend on domain-specific CT data and no further iterative refinement step was necessary, the proposed two-stage method achieves competitive results. The proposed method achieves a shared top ranking in the LoDoPaB-CT challenge and a first position with respect to the SSIM metric.
Authors: Raphael Baena, Lucas Drumetz, Vincent Gripon
Abstract: In classification, it is usual to observe that models trained on a given set of classes can generalize to previously unseen ones, suggesting the ability to learn beyond the initial task. This ability is often leveraged in the context of transfer learning where a pretrained model can be used to process new classes, with or without fine tuning. Surprisingly, there are a few papers looking at the theoretical roots beyond this phenomenon. In this work, we are interested in laying the foundations of such a theoretical framework for transferability between sets of classes. Namely, we establish a partially ordered set of subsets of classes. This tool allows to represent which subset of classes can generalize to others. In a more practical setting, we explore the ability of our framework to predict which subset of classes can lead to the best performance when testing on all of them. We also explore few-shot learning, where transfer is the golden standard. Our work contributes to better understanding of transfer mechanics and model generalization.
Authors: In-Gyu Lee, Jun-Young Oh, Hee-Jung Yu, Jae-Hwan Kim, Ki-Dong Eom, Ji-Hoon Jeong
Abstract: Recently, with increasing interest in pet healthcare, the demand for computer-aided diagnosis (CAD) systems in veterinary medicine has increased. The development of veterinary CAD has stagnated due to a lack of sufficient radiology data. To overcome the challenge, we propose a generative active learning framework based on a variational autoencoder. This approach aims to alleviate the scarcity of reliable data for CAD systems in veterinary medicine. This study utilizes datasets comprising cardiomegaly radiograph data. After removing annotations and standardizing images, we employed a framework for data augmentation, which consists of a data generation phase and a query phase for filtering the generated data. The experimental results revealed that as the data generated through this framework was added to the training data of the generative model, the frechet inception distance consistently decreased from 84.14 to 50.75 on the radiograph. Subsequently, when the generated data were incorporated into the training of the classification model, the false positive of the confusion matrix also improved from 0.16 to 0.66 on the radiograph. The proposed framework has the potential to address the challenges of data scarcity in medical CAD, contributing to its advancement.
Authors: Chuanyu Luo, Nuo Cheng, Ren Zhong, Haipeng Jiang, Wenyu Chen, Aoli Wang, Pu Li
Abstract: With the rapid advancement of hardware and software technologies, research in autonomous driving has seen significant growth. The prevailing framework for multi-sensor autonomous driving encompasses sensor installation, perception, path planning, decision-making, and motion control. At the perception phase, a common approach involves utilizing neural networks to infer 3D bounding box (Bbox) attributes from raw sensor data, including classification, size, and orientation. In this paper, we present a novel attribute and its corresponding algorithm: 3D object visibility. By incorporating multi-task learning, the introduction of this attribute, visibility, negligibly affects the model's effectiveness and efficiency. Our proposal of this attribute and its computational strategy aims to expand the capabilities for downstream tasks, thereby enhancing the safety and reliability of real-time autonomous driving in real-world scenarios.
Authors: Paul Doucet, Benjamin Estermann, Till Aczel, Roger Wattenhofer
Abstract: This study addresses the integration of diversity-based and uncertainty-based sampling strategies in active learning, particularly within the context of self-supervised pre-trained models. We introduce a straightforward heuristic called TCM that mitigates the cold start problem while maintaining strong performance across various data levels. By initially applying TypiClust for diversity sampling and subsequently transitioning to uncertainty sampling with Margin, our approach effectively combines the strengths of both strategies. Our experiments demonstrate that TCM consistently outperforms existing methods across various datasets in both low and high data regimes.
Authors: Yuta Ono, Till Aczel, Benjamin Estermann, Roger Wattenhofer
Abstract: Active learning is a machine learning paradigm designed to optimize model performance in a setting where labeled data is expensive to acquire. In this work, we propose a novel active learning method called SUPClust that seeks to identify points at the decision boundary between classes. By targeting these points, SUPClust aims to gather information that is most informative for refining the model's prediction of complex decision regions. We demonstrate experimentally that labeling these points leads to strong model performance. This improvement is observed even in scenarios characterized by strong class imbalance.
Authors: Joshua Springer, Gylfi {\TH}\'or Gu{\dh}mundsson, Marcel Kyas
Abstract: We propose a method for autonomous precision drone landing with fiducial markers and a gimbal-mounted, multi-payload camera with wide-angle, zoom, and IR sensors. The method has minimal data requirements; it depends primarily on the direction from the drone to the landing pad, enabling it to switch dynamically between the camera's different sensors and zoom factors, and minimizing auxiliary sensor requirements. It eliminates the need for data such as altitude above ground level, straight-line distance to the landing pad, fiducial marker size, and 6 DoF marker pose (of which the orientation is problematic). We leverage the zoom and wide-angle cameras, as well as visual April Tag fiducial markers to conduct successful precision landings from much longer distances than in previous work (168m horizontal distance, 102m altitude). We use two types of April Tags in the IR spectrum - active and passive - for precision landing both at daytime and nighttime, instead of simple IR beacons used in most previous work. The active IR landing pad is heated; the novel, passive one is unpowered, at ambient temperature, and depends on its high reflectivity and an IR differential between the ground and the sky. Finally, we propose a high-level control policy to manage initial search for the landing pad and subsequent searches if it is lost - not addressed in previous work. The method demonstrates successful landings with the landing skids at least touching the landing pad, achieving an average error of 0.19m. It also demonstrates successful recovery and landing when the landing pad is temporarily obscured.
Authors: Yubiao Yue, Zhenzhang Li
Abstract: Medical image classification is a very fundamental and crucial task in the field of computer vision. These years, CNN-based and Transformer-based models are widely used in classifying various medical images. Unfortunately, The limitation of CNNs in long-range modeling capabilities prevent them from effectively extracting fine-grained features in medical images , while Transformers are hampered by their quadratic computational complexity. Recent research has shown that the state space model (SSM) represented by Mamba can efficiently model long-range interactions while maintaining linear computational complexity. Inspired by this, we propose Vision Mamba for medical image classification (MedMamba). More specifically, we introduce a novel Conv-SSM module, which combines the local feature extraction ability of convolutional layers with the ability of SSM to capture long-range dependency. To demonstrate the potential of MedMamba, we conduct extensive experiments using three publicly available medical datasets with different imaging techniques (i.e., Kvasir (endoscopic images), FETAL_PLANES_DB (ultrasound images) and Covid19-Pneumonia-Normal Chest X-Ray (X-ray images)) and two private datasets built by ourselves. Experimental results show that the proposed MedMamba performs well in detecting lesions in various medical images. To the best of our knowledge, this is the first Vision Mamba tailored for medical image classification. The purpose of this work is to establish a new baseline for medical image classification tasks and provide valuable insights for the future development of more efficient and effective SSM-based artificial intelligence algorithms and application systems in the medical. Source code has been available at https://github.com/YubiaoYue/MedMamba.
Authors: Xiao Ma, Sumit Patidar, Iain Haughton, Stephen James
Abstract: This paper introduces Hierarchical Diffusion Policy (HDP), a hierarchical agent for multi-task robotic manipulation. HDP factorises a manipulation policy into a hierarchical structure: a high-level task-planning agent which predicts a distant next-best end-effector pose (NBP), and a low-level goal-conditioned diffusion policy which generates optimal motion trajectories. The factorised policy representation allows HDP to tackle both long-horizon task planning while generating fine-grained low-level actions. To generate context-aware motion trajectories while satisfying robot kinematics constraints, we present a novel kinematics-aware goal-conditioned control agent, Robot Kinematics Diffuser (RK-Diffuser). Specifically, RK-Diffuser learns to generate both the end-effector pose and joint position trajectories, and distill the accurate but kinematics-unaware end-effector pose diffuser to the kinematics-aware but less accurate joint position diffuser via differentiable kinematics. Empirically, we show that HDP achieves a significantly higher success rate than the state-of-the-art methods in both simulation and real-world.
Authors: Omar S. M. El Nahhas, Georg W\"olflein, Marta Ligero, Tim Lenz, Marko van Treeck, Firas Khader, Daniel Truhn, Jakob Nikolas Kather
Abstract: Deep Learning (DL) can predict biomarkers directly from digitized cancer histology in a weakly-supervised setting. Recently, the prediction of continuous biomarkers through regression-based DL has seen an increasing interest. Nonetheless, clinical decision making often requires a categorical outcome. Consequently, we developed a weakly-supervised joint multi-task Transformer architecture which has been trained and evaluated on four public patient cohorts for the prediction of two key predictive biomarkers, microsatellite instability (MSI) and homologous recombination deficiency (HRD), trained with auxiliary regression tasks related to the tumor microenvironment. Moreover, we perform a comprehensive benchmark of 16 approaches of task balancing for weakly-supervised joint multi-task learning in computational pathology. Using our novel approach, we improve over the state-of-the-art area under the receiver operating characteristic by +7.7% and +4.1%, as well as yielding better clustering of latent embeddings by +8% and +5% for the prediction of MSI and HRD in external cohorts, respectively.
Authors: Yanjie Ze, Gu Zhang, Kangning Zhang, Chenyuan Hu, Muhan Wang, Huazhe Xu
Abstract: Imitation learning provides an efficient way to teach robots dexterous skills; however, learning complex skills robustly and generalizablely usually consumes large amounts of human demonstrations. To tackle this challenging problem, we present 3D Diffusion Policy (DP3), a novel visual imitation learning approach that incorporates the power of 3D visual representations into diffusion policies, a class of conditional action generative models. The core design of DP3 is the utilization of a compact 3D visual representation, extracted from sparse point clouds with an efficient point encoder. In our experiments involving 72 simulation tasks, DP3 successfully handles most tasks with just 10 demonstrations and surpasses baselines with a 55.3% relative improvement. In 4 real robot tasks, DP3 demonstrates precise control with a high success rate of 85%, given only 40 demonstrations of each task, and shows excellent generalization abilities in diverse aspects, including space, viewpoint, appearance, and instance. Interestingly, in real robot experiments, DP3 rarely violates safety requirements, in contrast to baseline methods which frequently do, necessitating human intervention. Our extensive evaluation highlights the critical importance of 3D representations in real-world robot learning. Videos, code, and data are available on https://3d-diffusion-policy.github.io .
Authors: Juan Pablo Lagos, Esa Rahtu
Abstract: Holistic scene understanding is pivotal for the performance of autonomous machines. In this paper we propose a new end-to-end model for performing semantic segmentation and depth completion jointly. The vast majority of recent approaches have developed semantic segmentation and depth completion as independent tasks. Our approach relies on RGB and sparse depth as inputs to our model and produces a dense depth map and the corresponding semantic segmentation image. It consists of a feature extractor, a depth completion branch, a semantic segmentation branch and a joint branch which further processes semantic and depth information altogether. The experiments done on Virtual KITTI 2 dataset, demonstrate and provide further evidence, that combining both tasks, semantic segmentation and depth completion, in a multi-task network can effectively improve the performance of each task. Code is available at https://github.com/juanb09111/semantic depth.
Authors: XianFeng Han, Huixian Cheng, Hang Jiang, Dehong He, Guoqiang Xiao
Abstract: Fast and efficient semantic segmentation of large-scale LiDAR point clouds is a fundamental problem in autonomous driving. To achieve this goal, the existing point-based methods mainly choose to adopt Random Sampling strategy to process large-scale point clouds. However, our quantative and qualitative studies have found that Random Sampling may be less suitable for the autonomous driving scenario, since the LiDAR points follow an uneven or even long-tailed distribution across the space, which prevents the model from capturing sufficient information from points in different distance ranges and reduces the model's learning capability. To alleviate this problem, we propose a new Polar Cylinder Balanced Random Sampling method that enables the downsampled point clouds to maintain a more balanced distribution and improve the segmentation performance under different spatial distributions. In addition, a sampling consistency loss is introduced to further improve the segmentation performance and reduce the model's variance under different sampling methods. Extensive experiments confirm that our approach produces excellent performance on both SemanticKITTI and SemanticPOSS benchmarks, achieving a 2.8% and 4.0% improvement, respectively. The source code is available at https://github.com/huixiancheng/PCB-RandNet.
Authors: Juan Lagos, Esa Rahtu
Abstract: Understanding 3D environments semantically is pivotal in autonomous driving applications where multiple computer vision tasks are involved. Multi-task models provide different types of outputs for a given scene, yielding a more holistic representation while keeping the computational cost low. We propose a multi-task model for panoptic segmentation and depth completion using RGB images and sparse depth maps. Our model successfully predicts fully dense depth maps and performs semantic segmentation, instance segmentation, and panoptic segmentation for every input frame. Extensive experiments were done on the Virtual KITTI 2 dataset and we demonstrate that our model solves multiple tasks, without a significant increase in computational cost, while keeping high accuracy performance. Code is available at https://github.com/juanb09111/PanDepth.git
Authors: Rui Ding, Juntian Ye, Qifeng Gao, Feihu Xu, Yuping Duan
Abstract: Non-line-of-sight (NLOS) imaging aims to reconstruct the three-dimensional hidden scenes from the data measured in the line-of-sight, which uses photon time-of-flight information encoded in light after multiple diffuse reflections. The under-sampled scanning data can facilitate fast imaging. However, the resulting reconstruction problem becomes a serious ill-posed inverse problem, the solution of which is highly possibility to be degraded due to noises and distortions. In this paper, we propose novel NLOS reconstruction models based on curvature regularization, i.e., the object-domain curvature regularization model and the dual (signal and object)-domain curvature regularization model. In what follows, we develop efficient optimization algorithms relying on the alternating direction method of multipliers (ADMM) with the backtracking stepsize rule, for which all solvers can be implemented on GPUs. We evaluate the proposed algorithms on both synthetic and real datasets, which achieve state-of-the-art performance, especially in the compressed sensing setting. Based on GPU computing, our algorithm is the most effective among iterative methods, balancing reconstruction quality and computational time. All our codes and data are available at https://github.com/Duanlab123/CurvNLOS.
Authors: Jingzhe Ma, Dingqiang Ye, Chao Fan, Shiqi Yu
Abstract: As a kind of biometrics, the gait information of pedestrians has attracted widespread attention from both industry and academia since it can be acquired from long distances without the cooperation of targets. In recent literature, this line of research has brought exciting chances along with alarming challenges: On the positive side, gait recognition used for security applications such as suspect retrieval and safety checks is becoming more and more promising. On the negative side, the misuse of gait information may lead to privacy concerns, as lawbreakers can track subjects of interest using gait characteristics even under face-masked and clothes-changed scenarios. To handle this double-edged sword, we propose a gait attribute editing framework termed GaitEditor. It can perform various degrees of attribute edits on real gait sequences while maintaining the visual authenticity, respectively used for gait data augmentation and de-identification, thereby adaptively enhancing or degrading gait recognition performance according to users' intentions. Experimentally, we conduct a comprehensive evaluation under both gait recognition and anonymization protocols on three widely used gait benchmarks. Numerous results illustrate that the adaptable utilization of GaitEditor efficiently improves gait recognition performance and generates vivid visualizations with de-identification to protect human privacy. To the best of our knowledge, GaitEditor is the first framework capable of editing multiple gait attributes while simultaneously benefiting gait recognition and gait anonymization. The source code of GaitEditor will be available at https://github.com/ShiqiYu/OpenGait.
Authors: Guanfang Dong, Chenqiu Zhao, Xichen Pan, Anup Basu
Abstract: The goal of moving object segmentation is separating moving objects from stationary backgrounds in videos. One major challenge in this problem is how to develop a universal model for videos from various natural scenes since previous methods are often effective only in specific scenes. In this paper, we propose a method called Learning Temporal Distribution and Spatial Correlation (LTS) that has the potential to be a general solution for universal moving object segmentation. In the proposed approach, the distribution from temporal pixels is first learned by our Defect Iterative Distribution Learning (DIDL) network for a scene-independent segmentation. Notably, the DIDL network incorporates the use of an improved product distribution layer that we have newly derived. Then, the Stochastic Bayesian Refinement (SBR) Network, which learns the spatial correlation, is proposed to improve the binary mask generated by the DIDL network. Benefiting from the scene independence of the temporal distribution and the accuracy improvement resulting from the spatial correlation, the proposed approach performs well for almost all videos from diverse and complex natural scenes with fixed parameters. Comprehensive experiments on standard datasets including LASIESTA, CDNet2014, BMC, SBMI2015 and 128 real world videos demonstrate the superiority of proposed approach compared to state-of-the-art methods with or without the use of deep learning networks. To the best of our knowledge, this work has high potential to be a general solution for moving object segmentation in real world environments. The code and real-world videos can be found on GitHub https://github.com/guanfangdong/LTS-UniverisalMOS.
Authors: Yikun Liu, Jiangchao Yao, Ya Zhang, Yanfeng Wang, Weidi Xie
Abstract: In this paper, we consider the problem of composed image retrieval (CIR), it aims to train a model that can fuse multi-modal information, e.g., text and images, to accurately retrieve images that match the query, extending the user's expression ability. We make the following contributions: (i) we initiate a scalable pipeline to automatically construct datasets for training CIR model, by simply exploiting a large-scale dataset of image-text pairs, e.g., a subset of LAION-5B; (ii) we introduce a transformer-based adaptive aggregation model, TransAgg, which employs a simple yet efficient fusion mechanism, to adaptively combine information from diverse modalities; (iii) we conduct extensive ablation studies to investigate the usefulness of our proposed data construction procedure, and the effectiveness of core components in TransAgg; (iv) when evaluating on the publicly available benckmarks under the zero-shot scenario, i.e., training on the automatically constructed datasets, then directly conduct inference on target downstream datasets, e.g., CIRR and FashionIQ, our proposed approach either performs on par with or significantly outperforms the existing state-of-the-art (SOTA) models. Project page: https://code-kunkun.github.io/ZS-CIR/
Authors: Navid Rajabi, Jana Kosecka
Abstract: Large vision-and-language models (VLMs) trained to match images with text on large-scale datasets of image-text pairs have shown impressive generalization ability on several vision and language tasks. Several recent works, however, showed that these models lack fine-grained understanding, such as the ability to count and recognize verbs, attributes, or relationships. The focus of this work is to study the understanding of spatial relations. This has been tackled previously using image-text matching (e.g., Visual Spatial Reasoning benchmark) or visual question answering (e.g., GQA or VQAv2), both showing poor performance and a large gap compared to human performance. In this work, we show qualitatively (using explainability tools) and quantitatively (using object detectors) that the poor object localization "grounding" ability of the models is a contributing factor to the poor image-text matching performance. We propose an alternative fine-grained, compositional approach for recognizing and ranking spatial clauses that combines the evidence from grounding noun phrases corresponding to objects and their locations to compute the final rank of the spatial clause. We demonstrate the approach on representative VLMs (such as LXMERT, GPV, and MDETR) and compare and highlight their abilities to reason about spatial relationships.
Authors: Jonathan Xu, Amna Elmustafa, Liya Weldegebriel, Emnet Negash, Richard Lee, Chenlin Meng, Stefano Ermon, David Lobell
Abstract: Small farms contribute to a large share of the productive land in developing countries. In regions such as sub-Saharan Africa, where 80\% of farms are small (under 2 ha in size), the task of mapping smallholder cropland is an important part of tracking sustainability measures such as crop productivity. However, the visually diverse and nuanced appearance of small farms has limited the effectiveness of traditional approaches to cropland mapping. Here we introduce a new approach based on the detection of harvest piles characteristic of many smallholder systems throughout the world. We present HarvestNet, a dataset for mapping the presence of farms in the Ethiopian regions of Tigray and Amhara during 2020-2023, collected using expert knowledge and satellite images, totaling 7k hand-labeled images and 2k ground-collected labels. We also benchmark a set of baselines, including SOTA models in remote sensing, with our best models having around 80\% classification performance on hand labelled data and 90\% and 98\% accuracy on ground truth data for Tigray and Amhara, respectively. We also perform a visual comparison with a widely used pre-existing coverage map and show that our model detects an extra 56,621 hectares of cropland in Tigray. We conclude that remote sensing of harvest piles can contribute to more timely and accurate cropland assessments in food insecure regions. The dataset can be accessed through https://figshare.com/s/45a7b45556b90a9a11d2, while the code for the dataset and benchmarks is publicly available at https://github.com/jonxuxu/harvest-piles
URLs: https://figshare.com/s/45a7b45556b90a9a11d2,, https://github.com/jonxuxu/harvest-piles
Authors: Tianyi Song, Jiuxin Cao, Kun Wang, Bo Liu, Xiaofeng Zhang
Abstract: The excellent text-to-image synthesis capability of diffusion models has driven progress in synthesizing coherent visual stories. The current state-of-the-art method combines the features of historical captions, historical frames, and the current captions as conditions for generating the current frame. However, this method treats each historical frame and caption as the same contribution. It connects them in order with equal weights, ignoring that not all historical conditions are associated with the generation of the current frame. To address this issue, we propose Causal-Story. This model incorporates a local causal attention mechanism that considers the causal relationship between previous captions, frames, and current captions. By assigning weights based on this relationship, Causal-Story generates the current frame, thereby improving the global consistency of story generation. We evaluated our model on the PororoSV and FlintstonesSV datasets and obtained state-of-the-art FID scores, and the generated frames also demonstrate better storytelling in visuals.
Authors: Shengqi Liu, Zhuo Chen, Jingnan Gao, Yichao Yan, Wenhan Zhu, Jiangjing Lyu, Xiaokang Yang
Abstract: Texture editing is a crucial task in 3D modeling that allows users to automatically manipulate the surface materials of 3D models. However, the inherent complexity of 3D models and the ambiguous text description lead to the challenge in this task. To address this challenge, we propose ITEM3D, a \textbf{T}exture \textbf{E}diting \textbf{M}odel designed for automatic \textbf{3D} object editing according to the text \textbf{I}nstructions. Leveraging the diffusion models and the differentiable rendering, ITEM3D takes the rendered images as the bridge of text and 3D representation, and further optimizes the disentangled texture and environment map. Previous methods adopted the absolute editing direction namely score distillation sampling (SDS) as the optimization objective, which unfortunately results in the noisy appearance and text inconsistency. To solve the problem caused by the ambiguous text, we introduce a relative editing direction, an optimization objective defined by the noise difference between the source and target texts, to release the semantic ambiguity between the texts and images. Additionally, we gradually adjust the direction during optimization to further address the unexpected deviation in the texture domain. Qualitative and quantitative experiments show that our ITEM3D outperforms the state-of-the-art methods on various 3D objects. We also perform text-guided relighting to show explicit control over lighting. Our project page: https://shengqiliu1.github.io/ITEM3D.
Authors: Javier P\'erez de Frutos, Ragnhild Holden Helland, Shreya Desai, Line Cathrine Nymoen, Thomas Lang{\o}, Theodor Remman, Abhijit Sen
Abstract: Background: Dental caries diagnosis requires the manual inspection of diagnostic bitewing images of the patient, followed by a visual inspection and probing of the identified dental pieces with potential lesions. Yet the use of artificial intelligence, and in particular deep-learning, has the potential to aid in the diagnosis by providing a quick and informative analysis of the bitewing images. Methods: A dataset of 13,887 bitewings from the HUNT4 Oral Health Study were annotated individually by six different experts, and used to train three different object detection deep-learning architectures: RetinaNet (ResNet50), YOLOv5 (M size), and EfficientDet (D0 and D1 sizes). A consensus dataset of 197 images, annotated jointly by the same six dentist, was used for evaluation. A five-fold cross validation scheme was used to evaluate the performance of the AI models. Results: he trained models show an increase in average precision and F1-score, and decrease of false negative rate, with respect to the dental clinicians. When compared against the dental clinicians, the YOLOv5 model shows the largest improvement, reporting 0.647 mean average precision, 0.548 mean F1-score, and 0.149 mean false negative rate. Whereas the best annotators on each of these metrics reported 0.299, 0.495, and 0.164 respectively. Conclusion: Deep-learning models have shown the potential to assist dental professionals in the diagnosis of caries. Yet, the task remains challenging due to the artifacts natural to the bitewing images.
Authors: Boyang Zheng, Chumeng Liang, Xiaoyu Wu, Yan Liu
Abstract: Adversarial attacks on Latent Diffusion Model (LDM), the state-of-the-art image generative model, have been adopted as effective protection against malicious finetuning of LDM on unauthorized images. We show that these attacks add an extra error to the score function of adversarial examples predicted by LDM. LDM finetuned on these adversarial examples learns to lower the error by a bias, from which the model is attacked and predicts the score function with biases. Based on the dynamics, we propose to improve the adversarial attack on LDM by Attacking with Consistent score-function Errors (ACE). ACE unifies the pattern of the extra error added to the predicted score function. This induces the finetuned LDM to learn the same pattern as a bias in predicting the score function. We then introduce a well-crafted pattern to improve the attack. Our method outperforms state-of-the-art methods in adversarial attacks on LDM.
Authors: Cristiano Patr\'icio, Lu\'is F. Teixeira, Jo\~ao C. Neves
Abstract: Concept-based models naturally lend themselves to the development of inherently interpretable skin lesion diagnosis, as medical experts make decisions based on a set of visual patterns of the lesion. Nevertheless, the development of these models depends on the existence of concept-annotated datasets, whose availability is scarce due to the specialized knowledge and expertise required in the annotation process. In this work, we show that vision-language models can be used to alleviate the dependence on a large number of concept-annotated samples. In particular, we propose an embedding learning strategy to adapt CLIP to the downstream task of skin lesion classification using concept-based descriptions as textual embeddings. Our experiments reveal that vision-language models not only attain better accuracy when using concepts as textual embeddings, but also require a smaller number of concept-annotated samples to attain comparable performance to approaches specifically devised for automatic concept generation.
Authors: Zhen Wang, Xinyun Jiang, Jun Xiao, Tao Chen, Long Chen
Abstract: Explicit Caption Editing (ECE) -- refining reference image captions through a sequence of explicit edit operations (e.g., KEEP, DETELE) -- has raised significant attention due to its explainable and human-like nature. After training with carefully designed reference and ground-truth caption pairs, state-of-the-art ECE models exhibit limited generalization ability beyond the original training data distribution, i.e., they are tailored to refine content details only in in-domain samples but fail to correct errors in out-of-domain samples. To this end, we propose a new Diffusion-based Explicit Caption editing method: DECap. Specifically, we reformulate the ECE task as a denoising process under the diffusion mechanism, and introduce innovative edit-based noising and denoising processes. Thanks to this design, the noising process can help to eliminate the need for meticulous paired data selection by directly introducing word-level noises for training, learning diverse distribution over input reference caption. The denoising process involves the explicit predictions of edit operations and corresponding content words, refining reference captions through iterative step-wise editing. To further efficiently implement our diffusion process and improve the inference speed, DECap discards the prevalent multi-stage design and directly generates edit operations and content words simultaneously. Extensive ablations have demonstrated the strong generalization ability of DECap in various scenarios. More interestingly, it even shows great potential in improving the quality and controllability of caption generation.
Authors: Jiehong Lin, Lihua Liu, Dekun Lu, Kui Jia
Abstract: Zero-shot 6D object pose estimation involves the detection of novel objects with their 6D poses in cluttered scenes, presenting significant challenges for model generalizability. Fortunately, the recent Segment Anything Model (SAM) has showcased remarkable zero-shot transfer performance, which provides a promising solution to tackle this task. Motivated by this, we introduce SAM-6D, a novel framework designed to realize the task through two steps, including instance segmentation and pose estimation. Given the target objects, SAM-6D employs two dedicated sub-networks, namely Instance Segmentation Model (ISM) and Pose Estimation Model (PEM), to perform these steps on cluttered RGB-D images. ISM takes SAM as an advanced starting point to generate all possible object proposals and selectively preserves valid ones through meticulously crafted object matching scores in terms of semantics, appearance and geometry. By treating pose estimation as a partial-to-partial point matching problem, PEM performs a two-stage point matching process featuring a novel design of background tokens to construct dense 3D-3D correspondence, ultimately yielding the pose estimates. Without bells and whistles, SAM-6D outperforms the existing methods on the seven core datasets of the BOP Benchmark for both instance segmentation and pose estimation of novel objects.
Authors: Jia Li, Lijie Hu, Jingfeng Zhang, Tianhang Zheng, Hua Zhang, Di Wang
Abstract: In this paper, we address the limitations of existing text-to-image diffusion models in generating demographically fair results when given human-related descriptions. These models often struggle to disentangle the target language context from sociocultural biases, resulting in biased image generation. To overcome this challenge, we propose Fair Mapping, a flexible, model-agnostic, and lightweight approach that modifies a pre-trained text-to-image diffusion model by controlling the prompt to achieve fair image generation. One key advantage of our approach is its high efficiency. It only requires updating an additional linear network with few parameters at a low computational cost. By developing a linear network that maps conditioning embeddings into a debiased space, we enable the generation of relatively balanced demographic results based on the specified text condition. With comprehensive experiments on face image generation, we show that our method significantly improves image generation fairness with almost the same image quality compared to conventional diffusion models when prompted with descriptions related to humans. By effectively addressing the issue of implicit language bias, our method produces more fair and diverse image outputs.
Authors: Guozhang Li, Xinpeng Ding, De Cheng, Jie Li, Nannan Wang, Xinbo Gao
Abstract: Early weakly supervised video grounding (WSVG) methods often struggle with incomplete boundary detection due to the absence of temporal boundary annotations. To bridge the gap between video-level and boundary-level annotation, explicit-supervision methods, i.e., generating pseudo-temporal boundaries for training, have achieved great success. However, data augmentations in these methods might disrupt critical temporal information, yielding poor pseudo boundaries. In this paper, we propose a new perspective that maintains the integrity of the original temporal content while introducing more valuable information for expanding the incomplete boundaries. To this end, we propose EtC (Expand then Clarify), first use the additional information to expand the initial incomplete pseudo boundaries, and subsequently refine these expanded ones to achieve precise boundaries. Motivated by video continuity, i.e., visual similarity across adjacent frames, we use powerful multimodal large language models (MLLMs) to annotate each frame within initial pseudo boundaries, yielding more comprehensive descriptions for expanded boundaries. To further clarify the noise of expanded boundaries, we combine mutual learning with a tailored proposal-level contrastive objective to use a learnable approach to harmonize a balance between incomplete yet clean (initial) and comprehensive yet noisy (expanded) boundaries for more precise ones. Experiments demonstrate the superiority of our method on two challenging WSVG datasets.
Authors: Chenhuan Li, Meihua Xiao, zehuan li, Fangping Chen, Shanshan Qiao, Dingli Wang, Mengxi Gao, Siyi Zhang
Abstract: Recently, vision transformers have performed well in various computer vision tasks, including voxel 3D reconstruction. However, the windows of the vision transformer are not multi-scale, and there is no connection between the windows, which limits the accuracy of voxel 3D reconstruction. Therefore, we propose a voxel 3D reconstruction network based on shifted window attention. To the best of our knowledge, this is the first work to apply shifted window attention to voxel 3D reconstruction. Experimental results on ShapeNet verify our method achieves SOTA accuracy in single-view reconstruction.
Authors: Wonjoon Chang, Dahee Kwon, Jaesik Choi
Abstract: Understanding intermediate representations of the concepts learned by deep learning classifiers is indispensable for interpreting general model behaviors. Existing approaches to reveal learned concepts often rely on human supervision, such as pre-defined concept sets or segmentation processes. In this paper, we propose a novel unsupervised method for discovering distributed representations of concepts by selecting a principal subset of neurons. Our empirical findings demonstrate that instances with similar neuron activation states tend to share coherent concepts. Based on the observations, the proposed method selects principal neurons that construct an interpretable region, namely a Relaxed Decision Region (RDR), encompassing instances with coherent concepts in the feature space. It can be utilized to identify unlabeled subclasses within data and to detect the causes of misclassifications. Furthermore, the applicability of our method across various layers discloses distinct distributed representations over the layers, which provides deeper insights into the internal mechanisms of the deep learning model.
Authors: Peng Zheng, Dehong Gao, Deng-Ping Fan, Li Liu, Jorma Laaksonen, Wanli Ouyang, Nicu Sebe
Abstract: We introduce a novel bilateral reference framework (BiRefNet) for high-resolution dichotomous image segmentation (DIS). It comprises two essential components: the localization module (LM) and the reconstruction module (RM) with our proposed bilateral reference (BiRef). The LM aids in object localization using global semantic information. Within the RM, we utilize BiRef for the reconstruction process, where hierarchical patches of images provide the source reference and gradient maps serve as the target reference. These components collaborate to generate the final predicted maps. We also introduce auxiliary gradient supervision to enhance focus on regions with finer details. Furthermore, we outline practical training strategies tailored for DIS to improve map quality and training process. To validate the general applicability of our approach, we conduct extensive experiments on four tasks to evince that BiRefNet exhibits remarkable performance, outperforming task-specific cutting-edge methods across all benchmarks. Our codes are available at https://github.com/ZhengPeng7/BiRefNet.
Authors: Gyutae Hwang, Sang Jun Lee
Abstract: Human health can be critically affected by cardiovascular diseases, such as hypertension, arrhythmias, and stroke. Heart rate and blood pressure are important biometric information for the monitoring of cardiovascular system and early diagnosis of cardiovascular diseases. Existing methods for estimating the heart rate are based on electrocardiography and photoplethyomography, which require contacting the sensor to the skin surface. Moreover, catheter and cuff-based methods for measuring blood pressure cause inconvenience and have limited applicability. Therefore, in this thesis, we propose a vision-based method for estimating the heart rate and blood pressure. This thesis proposes a 2-stage deep learning framework consisting of a dual remote photoplethysmography network (DRP-Net) and bounded blood pressure network (BBP-Net). In the first stage, DRP-Net infers remote photoplethysmography (rPPG) signals for the acral and facial regions, and these phase-shifted rPPG signals are utilized to estimate the heart rate. In the second stage, BBP-Net integrates temporal features and analyzes phase discrepancy between the acral and facial rPPG signals to estimate SBP and DBP values. To improve the accuracy of estimating the heart rate, we employed a data augmentation method based on a frame interpolation model. Moreover, we designed BBP-Net to infer blood pressure within a predefined range by incorporating a scaled sigmoid function. Our method resulted in estimating the heart rate with the mean absolute error (MAE) of 1.78 BPM, reducing the MAE by 34.31 % compared to the recent method, on the MMSE-HR dataset. The MAE for estimating the systolic blood pressure (SBP) and diastolic blood pressure (DBP) were 10.19 mmHg and 7.09 mmHg. On the V4V dataset, the MAE for the heart rate, SBP, and DBP were 3.83 BPM, 13.64 mmHg, and 9.4 mmHg, respectively.
Authors: Yatong Bai, Utsav Garg, Apaar Shanker, Haoming Zhang, Samyak Parajuli, Erhan Bas, Isidora Filipovic, Amelia N. Chu, Eugenia D Fomitcheva, Elliot Branson, Aerin Kim, Somayeh Sojoudi, Kyunghyun Cho
Abstract: Vision and vision-language applications of neural networks, such as image classification and captioning, rely on large-scale annotated datasets that require non-trivial data-collecting processes. This time-consuming endeavor hinders the emergence of large-scale datasets, limiting researchers and practitioners to a small number of choices. Therefore, we seek more efficient ways to collect and annotate images. Previous initiatives have gathered captions from HTML alt-texts and crawled social media postings, but these data sources suffer from noise, sparsity, or subjectivity. For this reason, we turn to commercial shopping websites whose data meet three criteria: cleanliness, informativeness, and fluency. We introduce the Let's Go Shopping (LGS) dataset, a large-scale public dataset with 15 million image-caption pairs from publicly available e-commerce websites. When compared with existing general-domain datasets, the LGS images focus on the foreground object and have less complex backgrounds. Our experiments on LGS show that the classifiers trained on existing benchmark datasets do not readily generalize to e-commerce data, while specific self-supervised visual feature extractors can better generalize. Furthermore, LGS's high-quality e-commerce-focused images and bimodal nature make it advantageous for vision-language bi-modal tasks: LGS enables image-captioning models to generate richer captions and helps text-to-image generation models achieve e-commerce style transfer.
Authors: Baoxiong Jia, Yixin Chen, Huangyue Yu, Yan Wang, Xuesong Niu, Tengyu Liu, Qing Li, Siyuan Huang
Abstract: 3D vision-language grounding, which focuses on aligning language with the 3D physical environment, stands as a cornerstone in the development of embodied agents. In comparison to recent advancements in the 2D domain, grounding language in 3D scenes faces several significant challenges: (i) the inherent complexity of 3D scenes due to the diverse object configurations, their rich attributes, and intricate relationships; (ii) the scarcity of paired 3D vision-language data to support grounded learning; and (iii) the absence of a unified learning framework to distill knowledge from grounded 3D data. In this work, we aim to address these three major challenges in 3D vision-language by examining the potential of systematically upscaling 3D vision-language learning in indoor environments. We introduce the first million-scale 3D vision-language dataset, SceneVerse, encompassing about 68K 3D indoor scenes and comprising 2.5M vision-language pairs derived from both human annotations and our scalable scene-graph-based generation approach. We demonstrate that this scaling allows for a unified pre-training framework, Grounded Pre-training for Scenes (GPS), for 3D vision-language learning. Through extensive experiments, we showcase the effectiveness of GPS by achieving state-of-the-art performance on all existing 3D visual grounding benchmarks. The vast potential of SceneVerse and GPS is unveiled through zero-shot transfer experiments in the challenging 3D vision-language tasks. Project website: https://scene-verse.github.io.
Authors: Haibi Wang, Weifeng Ge
Abstract: With the breakthrough of multi-modal large language models, answering complex visual questions that demand advanced reasoning abilities and world knowledge has become a much more important testbed for developing AI models than ever. However, equipping AI models with robust cross-modality reasoning ability remains challenging since the cognition scheme of humans has not been understood systematically. In this paper, we believe that if we can collect visual clues in the given image as much as possible, we will recognize the image more accurately, understand the question better, recall relevant knowledge more easily, and finally reason out the answer. We discover these rich visual clues by mining question-answer pairs in images and sending them into multi-modal large language models as prompts. We call the proposed method Q&A Prompts. Specifically, we first use the image-answer pairs and the corresponding questions in the training set as inputs and outputs to train a visual question generation model. Then, we use an image tagging model to identify various instances and send packaged image-tag pairs into the visual question generation model to generate relevant questions with the extracted image tags as answers. Finally, we encode these generated question-answer pairs as prompts with a visual-aware prompting module and send them into pre-trained multi-modal large language models to reason out the final answers. Experimental results show that, compared with state-of-the-art methods, our Q&A Prompts achieves substantial improvements on the challenging visual question answering datasets requiring reasoning over diverse world knowledge, such as OK-VQA and A-OKVQA.
Authors: Shuchang Ye, Mingyuan Meng, Mingjian Li, Dagan Feng, Jinman Kim
Abstract: With increasing reliance on medical imaging in clinical practices, automated report generation from medical images is in great demand. Existing report generation methods typically adopt an encoder-decoder deep learning framework to build a uni-directional image-to-report mapping. However, such a framework ignores the bi-directional mutual associations between images and reports, thus incurring difficulties in associating the intrinsic medical meanings between them. Recent generative representation learning methods have demonstrated the benefits of dual-modal learning from both image and text modalities. However, these methods exhibit two major drawbacks for medical report generation: 1) they tend to capture morphological information and have difficulties in capturing subtle pathological semantic information, and 2) they predict masked text rely on both unmasked images and text, inevitably degrading performance when inference is based solely on images. In this study, we propose a new report generation framework with dual-modal dynamic traceback learning (DTrace) to overcome the two identified drawbacks and enable dual-modal learning for medical report generation. To achieve this, our DTrace introduces a traceback mechanism to control the semantic validity of generated content via self-assessment. Further, our DTrace introduces a dynamic learning strategy to adapt to various proportions of image and text input, enabling report generation without reliance on textual input during inference. Extensive experiments on two well-benchmarked datasets (IU-Xray and MIMIC-CXR) show that our DTrace outperforms state-of-the-art medical report generation methods.
Authors: Nishchal Sapkota, Yejia Zhang, Sirui Li, Peixian Liang, Zhuo Zhao, Jingjing Zhang, Xiaomin Zha, Yiru Zhou, Yunxia Cao, Danny Z Chen
Abstract: Male infertility accounts for about one-third of global infertility cases. Manual assessment of sperm abnormalities through head morphology analysis encounters issues of observer variability and diagnostic discrepancies among experts. Its alternative, Computer-Assisted Semen Analysis (CASA), suffers from low-quality sperm images, small datasets, and noisy class labels. We propose a new approach for sperm head morphology classification, called SHMC-Net, which uses segmentation masks of sperm heads to guide the morphology classification of sperm images. SHMC-Net generates reliable segmentation masks using image priors, refines object boundaries with an efficient graph-based method, and trains an image network with sperm head crops and a mask network with the corresponding masks. In the intermediate stages of the networks, image and mask features are fused with a fusion scheme to better learn morphological features. To handle noisy class labels and regularize training on small datasets, SHMC-Net applies Soft Mixup to combine mixup augmentation and a loss function. We achieve state-of-the-art results on SCIAN and HuSHeM datasets, outperforming methods that use additional pre-training or costly ensembling techniques.
Authors: Yuyue Zhou, Banafshe Felfeliyan, Shrimanti Ghosh, Jessica Knight, Fatima Alves-Pereira, Christopher Keen, Jessica K\"upper, Abhilash Rakkunedeth Hareendranathan, Jacob L. Jaremko
Abstract: Conventional deep learning models deal with images one-by-one, requiring costly and time-consuming expert labeling in the field of medical imaging, and domain-specific restriction limits model generalizability. Visual in-context learning (ICL) is a new and exciting area of research in computer vision. Unlike conventional deep learning, ICL emphasizes the model's ability to adapt to new tasks based on given examples quickly. Inspired by MAE-VQGAN, we proposed a new simple visual ICL method called SimICL, combining visual ICL pairing images with masked image modeling (MIM) designed for self-supervised learning. We validated our method on bony structures segmentation in a wrist ultrasound (US) dataset with limited annotations, where the clinical objective was to segment bony structures to help with further fracture detection. We used a test set containing 3822 images from 18 patients for bony region segmentation. SimICL achieved an remarkably high Dice coeffient (DC) of 0.96 and Jaccard Index (IoU) of 0.92, surpassing state-of-the-art segmentation and visual ICL models (a maximum DC 0.86 and IoU 0.76), with SimICL DC and IoU increasing up to 0.10 and 0.16. This remarkably high agreement with limited manual annotations indicates SimICL could be used for training AI models even on small US datasets. This could dramatically decrease the human expert time required for image labeling compared to conventional approaches, and enhance the real-world use of AI assistance in US image analysis.
Authors: Yanqi Qiao, Dazhuang Liu, Rui Wang, Kaitai Liang
Abstract: While convolutional neural networks (CNNs) have achieved success in computer vision tasks, it is vulnerable to backdoor attacks. Such attacks could mislead the victim model to make attacker-chosen prediction with a specific trigger pattern. Until now, the trigger injection of existing attacks is mainly limited to spatial domain. Recent works take advantage of perceptual properties of planting specific patterns in the frequency domain, which only reflect indistinguishable pixel-wise perturbations in pixel domain. However, in the black-box setup, the inaccessibility of training process often renders more complex trigger designs. Existing frequency attacks simply handcraft the magnitude of spectrum, introducing anomaly frequency disparities between clean and poisoned data and taking risks of being removed by image processing operations (such as lossy compression and filtering). In this paper, we propose a robust low-frequency black-box backdoor attack (LFBA), which minimally perturbs low-frequency components of frequency spectrum and maintains the perceptual similarity in spatial space simultaneously. The key insight of our attack restrict the search for the optimal trigger to low-frequency region that can achieve high attack effectiveness, robustness against image transformation defenses and stealthiness in dual space. We utilize simulated annealing (SA), a form of evolutionary algorithm, to optimize the properties of frequency trigger including the number of manipulated frequency bands and the perturbation of each frequency component, without relying on the knowledge from the victim classifier. Extensive experiments on real-world datasets verify the effectiveness and robustness of LFBA against image processing operations and the state-of-the-art backdoor defenses, as well as its inherent stealthiness in both spatial and frequency space, making it resilient against frequency inspection.
Authors: Huan Ni, Yubin Zhao, Haiyan Guan, Cheng Jiang, Yongshi Jie, Xing Wang, Yiyang Shen
Abstract: Large-scale high-resolution land cover classification is a prerequisite for constructing Earth system models and addressing ecological and resource issues. Advancements in satellite sensor technology have led to an improvement in spatial resolution and wider coverage areas. Nevertheless, the lack of high-resolution labeled data is still a challenge, hindering the largescale application of land cover classification methods. In this paper, we propose a Transformerbased weakly supervised method for cross-resolution land cover classification using outdated data. First, to capture long-range dependencies without missing the fine-grained details of objects, we propose a U-Net-like Transformer based on a reverse difference mechanism (RDM) using dynamic sparse attention. Second, we propose an anti-noise loss calculation (ANLC) module based on optimal transport (OT). Anti-noise loss calculation identifies confident areas (CA) and vague areas (VA) based on the OT matrix, which relieves the impact of noises in outdated land cover products. By introducing a weakly supervised loss with weights and employing unsupervised loss, the RDM-based U-Net-like Transformer was trained. Remote sensing images with 1 m resolution and the corresponding ground-truths of six states in the United States were employed to validate the performance of the proposed method. The experiments utilized outdated land cover products with 30 m resolution from 2013 as training labels, and produced land cover maps with 1 m resolution from 2017. The results show the superiority of the proposed method compared to state-of-the-art methods. The code is available at https://github.com/yu-ni1989/ANLC-Former.
Authors: Zekun Qi, Runpei Dong, Shaochen Zhang, Haoran Geng, Chunrui Han, Zheng Ge, He Wang, Li Yi, Kaisheng Ma
Abstract: This paper presents ShapeLLM, the first 3D Multimodal Large Language Model (LLM) designed for embodied interaction, exploring a universal 3D object understanding with 3D point clouds and languages. ShapeLLM is built upon an improved 3D encoder by extending ReCon to ReCon++ that benefits from multi-view image distillation for enhanced geometry understanding. By utilizing ReCon++ as the 3D point cloud input encoder for LLMs, ShapeLLM is trained on constructed instruction-following data and tested on our newly human-curated evaluation benchmark, 3D MM-Vet. ReCon++ and ShapeLLM achieve state-of-the-art performance in 3D geometry understanding and language-unified 3D interaction tasks, such as embodied visual grounding.
Authors: Weilin Wan, Weizhong Zhang, Cheng Jin
Abstract: Out-of-distribution detection (OOD) is a crucial technique for deploying machine learning models in the real world to handle the unseen scenarios. In this paper, we first propose a simple yet effective Neural Activation Prior (NAP) for OOD detection. Our neural activation prior is based on a key observation that, for a channel before the global pooling layer of a fully trained neural network, the probability of a few neurons being activated with a large response by an in-distribution (ID) sample is significantly higher than that by an OOD sample. An intuitive explanation is that for a model fully trained on ID dataset, each channel would play a role in detecting a certain pattern in the ID dataset, and a few neurons can be activated with a large response when the pattern is detected in an input sample. Then, a new scoring function based on this prior is proposed to highlight the role of these strongly activated neurons in OOD detection. Our approach is plug-and-play and does not lead to any performance degradation on ID data classification and requires no extra training or statistics from training or external datasets. Notice that previous methods primarily rely on post-global-pooling features of the neural networks, while the within-channel distribution information we leverage would be discarded by the global pooling operator. Consequently, our method is orthogonal to existing approaches and can be effectively combined with them in various applications. Experimental results show that our method achieves the state-of-the-art performance on CIFAR benchmark and ImageNet dataset, which demonstrates the power of the proposed prior. Finally, we extend our method to Transformers and the experimental findings indicate that NAP can also significantly enhance the performance of OOD detection on Transformers, thereby demonstrating the broad applicability of this prior knowledge.
Authors: Xuzhe Zhang, Elsa D. Angelini, Eric A. Hoffman, Karol E. Watson, Benjamin M. Smith, R. Graham Barr, Andrew F. Laine
Abstract: Robust quantification of pulmonary emphysema on computed tomography (CT) remains challenging for large-scale research studies that involve scans from different scanner types and for translation to clinical scans. Existing studies have explored several directions to tackle this challenge, including density correction, noise filtering, regression, hidden Markov measure field (HMMF) model-based segmentation, and volume-adjusted lung density. Despite some promising results, previous studies either required a tedious workflow or limited opportunities for downstream emphysema subtyping, limiting efficient adaptation on a large-scale study. To alleviate this dilemma, we developed an end-to-end deep learning framework based on an existing HMMF segmentation framework. We first demonstrate that a regular UNet cannot replicate the existing HMMF results because of the lack of scanner priors. We then design a novel domain attention block to fuse image feature with quantitative scanner priors which significantly improves the results.
Authors: Rui Gong, Weide Liu, Zaiwang Gu, Xulei Yang, Jun Cheng
Abstract: Geometric knowledge has been shown to be beneficial for the stereo matching task. However, prior attempts to integrate geometric insights into stereo matching algorithms have largely focused on geometric knowledge from single images while crucial cross-view factors such as occlusion and matching uniqueness have been overlooked. To address this gap, we propose a novel Intra-view and Cross-view Geometric knowledge learning Network (ICGNet), specifically crafted to assimilate both intra-view and cross-view geometric knowledge. ICGNet harnesses the power of interest points to serve as a channel for intra-view geometric understanding. Simultaneously, it employs the correspondences among these points to capture cross-view geometric relationships. This dual incorporation empowers the proposed ICGNet to leverage both intra-view and cross-view geometric knowledge in its learning process, substantially improving its ability to estimate disparities. Our extensive experiments demonstrate the superiority of the ICGNet over contemporary leading models.
Authors: Haonan Wang, Qixiang Zhang, Yi Li, Xiaomeng Li
Abstract: Semi-supervised semantic segmentation (SSSS) has been proposed to alleviate the burden of time-consuming pixel-level manual labeling, which leverages limited labeled data along with larger amounts of unlabeled data. Current state-of-the-art methods train the labeled data with ground truths and unlabeled data with pseudo labels. However, the two training flows are separate, which allows labeled data to dominate the training process, resulting in low-quality pseudo labels and, consequently, sub-optimal results. To alleviate this issue, we present AllSpark, which reborns the labeled features from unlabeled ones with the channel-wise cross-attention mechanism. We further introduce a Semantic Memory along with a Channel Semantic Grouping strategy to ensure that unlabeled features adequately represent labeled features. The AllSpark shed new light on the architecture level designs of SSSS rather than framework level, which avoids increasingly complicated training pipeline designs. It can also be regarded as a flexible bottleneck module that can be seamlessly integrated into a general transformer-based segmentation model. The proposed AllSpark outperforms existing methods across all evaluation protocols on Pascal, Cityscapes and COCO benchmarks without bells-and-whistles. Code and model weights are available at: https://github.com/xmed-lab/AllSpark.
Authors: Xuweiyi Chen, Tian Xia, Sihan Xu
Abstract: Video Diffusion Models have been developed for video generation, usually integrating text and image conditioning to enhance control over the generated content. Despite the progress, ensuring consistency across frames remains a challenge, particularly when using text prompts as control conditions. To address this problem, we introduce UniCtrl, a novel, plug-and-play method that is universally applicable to improve the spatiotemporal consistency and motion diversity of videos generated by text-to-video models without additional training. UniCtrl ensures semantic consistency across different frames through cross-frame self-attention control, and meanwhile, enhances the motion quality and spatiotemporal consistency through motion injection and spatiotemporal synchronization. Our experimental results demonstrate UniCtrl's efficacy in enhancing various text-to-video models, confirming its effectiveness and universality.
Authors: Cong Ma, Lei Qiao, Chengkai Zhu, Kai Liu, Zelong Kong, Qing Li, Xueqi Zhou, Yuheng Kan, Wei Wu
Abstract: Vehicle-to-everything (V2X) is a popular topic in the field of Autonomous Driving in recent years. Vehicle-infrastructure cooperation (VIC) becomes one of the important research area. Due to the complexity of traffic conditions such as blind spots and occlusion, it greatly limits the perception capabilities of single-view roadside sensing systems. To further enhance the accuracy of roadside perception and provide better information to the vehicle side, in this paper, we constructed holographic intersections with various layouts to build a large-scale multi-sensor holographic vehicle-infrastructure cooperation dataset, called HoloVIC. Our dataset includes 3 different types of sensors (Camera, Lidar, Fisheye) and employs 4 sensor-layouts based on the different intersections. Each intersection is equipped with 6-18 sensors to capture synchronous data. While autonomous vehicles pass through these intersections for collecting VIC data. HoloVIC contains in total on 100k+ synchronous frames from different sensors. Additionally, we annotated 3D bounding boxes based on Camera, Fisheye, and Lidar. We also associate the IDs of the same objects across different devices and consecutive frames in sequence. Based on HoloVIC, we formulated four tasks to facilitate the development of related research. We also provide benchmarks for these tasks.
Authors: Zheng Li, Xiang Li, Xinyi Fu, Xing Zhang, Weiqiang Wang, Shuo Chen, Jian Yang
Abstract: Prompt learning has emerged as a valuable technique in enhancing vision-language models (VLMs) such as CLIP for downstream tasks in specific domains. Existing work mainly focuses on designing various learning forms of prompts, neglecting the potential of prompts as effective distillers for learning from larger teacher models. In this paper, we introduce an unsupervised domain prompt distillation framework, which aims to transfer the knowledge of a larger teacher model to a lightweight target model through prompt-driven imitation using unlabeled domain images. Specifically, our framework consists of two distinct stages. In the initial stage, we pre-train a large CLIP teacher model using domain (few-shot) labels. After pre-training, we leverage the unique decoupled-modality characteristics of CLIP by pre-computing and storing the text features as class vectors only once through the teacher text encoder. In the subsequent stage, the stored class vectors are shared across teacher and student image encoders for calculating the predicted logits. Further, we align the logits of both the teacher and student models via KL divergence, encouraging the student image encoder to generate similar probability distributions to the teacher through the learnable prompts. The proposed prompt distillation process eliminates the reliance on labeled data, enabling the algorithm to leverage a vast amount of unlabeled images within the domain. Finally, the well-trained student image encoders and pre-stored text features (class vectors) are utilized for inference. To our best knowledge, we are the first to (1) perform unsupervised domain-specific prompt-driven knowledge distillation for CLIP, and (2) establish a practical pre-storing mechanism of text features as shared class vectors between teacher and student. Extensive experiments on 11 datasets demonstrate the effectiveness of our method.
Authors: Xijia Tao, Shuai Zhong, Lei Li, Qi Liu, Lingpeng Kong
Abstract: There has been an increasing interest in the alignment of large language models (LLMs) with human values. However, the safety issues of their integration with a vision module, or vision language models (VLMs), remain relatively underexplored. In this paper, we propose a novel jailbreaking attack against VLMs, aiming to bypass their safety barrier when a user inputs harmful instructions. A scenario where our poisoned (image, text) data pairs are included in the training data is assumed. By replacing the original textual captions with malicious jailbreak prompts, our method can perform jailbreak attacks with the poisoned images. Moreover, we analyze the effect of poison ratios and positions of trainable parameters on our attack's success rate. For evaluation, we design two metrics to quantify the success rate and the stealthiness of our attack. Together with a list of curated harmful instructions, a benchmark for measuring attack efficacy is provided. We demonstrate the efficacy of our attack by comparing it with baseline methods.
Authors: Ruizhuo Song, Beiming Yuan
Abstract: Abstract reasoning problems pose significant challenges to artificial intelligence algorithms, demanding cognitive capabilities beyond those required for perception tasks. This study introduces the Triple-CFN approach to tackle the Bongard-Logo problem, achieving notable reasoning accuracy by implicitly reorganizing the concept space of conflicting instances. Additionally, the Triple-CFN paradigm proves effective for the RPM problem with necessary modifications, yielding competitive results. To further enhance performance on the RPM issue, we develop the Meta Triple-CFN network, which explicitly structures the problem space while maintaining interpretability on progressive patterns. The success of Meta Triple-CFN is attributed to its paradigm of modeling the conceptual space, equivalent to normalizing reasoning information. Based on this ideology, we introduce the Re-space layer, enhancing the performance of both Meta Triple-CFN and Triple-CFN. This paper aims to contribute to advancements in machine intelligence by exploring innovative network designs for addressing abstract reasoning problems, paving the way for further breakthroughs in this domain.
Authors: Erik Bauer, Barnabas Gavin Cangan, Robert K. Katzschmann
Abstract: In a future with autonomous robots, visual and spatial perception is of utmost importance for robotic systems. Particularly for aerial robotics, there are many applications where utilizing visual perception is necessary for any real-world scenarios. Robotic aerial grasping using drones promises fast pick-and-place solutions with a large increase in mobility over other robotic solutions. Utilizing Mask R-CNN scene segmentation (detectron2), we propose a vision-based system for autonomous rapid aerial grasping which does not rely on markers for object localization and does not require the appearance of the object to be previously known. Combining segmented images with spatial information from a depth camera, we generate a dense point cloud of the detected objects and perform geometry-based grasp planning to determine grasping points on the objects. In real-world experiments on a dynamically grasping aerial platform, we show that our system can replicate the performance of a motion capture system for object localization up to 94.5 % of the baseline grasping success rate. With our results, we show the first use of geometry-based grasping techniques with a flying platform and aim to increase the autonomy of existing aerial manipulation platforms, bringing them further towards real-world applications in warehouses and similar environments.
Authors: Junwei Zheng, Jiaming Zhang, Kailun Yang, Kunyu Peng, Rainer Stiefelhagen
Abstract: People with Visual Impairments (PVI) typically recognize objects through haptic perception. Knowing objects and materials before touching is desired by the target users but under-explored in the field of human-centered robotics. To fill this gap, in this work, a wearable vision-based robotic system, MateRobot, is established for PVI to recognize materials and object categories beforehand. To address the computational constraints of mobile platforms, we propose a lightweight yet accurate model MateViT to perform pixel-wise semantic segmentation, simultaneously recognizing both objects and materials. Our methods achieve respective 40.2% and 51.1% of mIoU on COCOStuff-10K and DMS datasets, surpassing the previous method with +5.7% and +7.0% gains. Moreover, on the field test with participants, our wearable system reaches a score of 28 in the NASA-Task Load Index, indicating low cognitive demands and ease of use. Our MateRobot demonstrates the feasibility of recognizing material property through visual cues and offers a promising step towards improving the functionality of wearable robots for PVI. The source code has been made publicly available at https://junweizheng93.github.io/publications/MATERobot/MATERobot.html.
URLs: https://junweizheng93.github.io/publications/MATERobot/MATERobot.html.
Authors: Vithursan Thangarasa, Shreyas Saxena, Abhay Gupta, Sean Lie
Abstract: Recent research has focused on weight sparsity in neural network training to reduce FLOPs, aiming for improved efficiency (test accuracy w.r.t training FLOPs). However, sparse weight training often sacrifices accuracy, requiring extended training schedules to attain the accuracy of dense models. In contrast, our approach, Sparse Iso-FLOP Transformations (Sparse-IFT), uses sparsity to improve accuracy while maintaining dense model FLOPs. Using a single hyperparameter (i.e., sparsity level), Sparse-IFTs efficiently replace dense layers, expanding the search space for optimal sparse masks. In addition, dynamic sparse training with Sparse-IFT models effectively navigates this larger sparse mask-weight space, which is evidenced by a spectral analysis using Ramanujan graph properties. Our study reveals a robust correlation among mask topology, weights, and final performance. Notably, without adjusting hyperparameters, replacing dense layers with Sparse-IFT yields significant improvements, such as a +3.5% boost for ResNet-18 on ImageNet and +0.9% for GPT-3 Small on the Open LLM leaderboard. To our knowledge, this is the first work to demonstrate the use of sparsity for improving the accuracy of dense models through a simple-to-use set of sparse transformations. Code is available at: https://github.com/CerebrasResearch/Sparse-IFT.
Authors: Emiel Hoogeboom, Eirikur Agustsson, Fabian Mentzer, Luca Versari, George Toderici, Lucas Theis
Abstract: Despite the tremendous success of diffusion generative models in text-to-image generation, replicating this success in the domain of image compression has proven difficult. In this paper, we demonstrate that diffusion can significantly improve perceptual quality at a given bit-rate, outperforming state-of-the-art approaches PO-ELIC and HiFiC as measured by FID score. This is achieved using a simple but theoretically motivated two-stage approach combining an autoencoder targeting MSE followed by a further score-based decoder. However, as we will show, implementation details matter and the optimal design decisions can differ greatly from typical text-to-image models.
Authors: Ling Yang, Ye Tian, Minkai Xu, Zhongyi Liu, Shenda Hong, Wei Qu, Wentao Zhang, Bin Cui, Muhan Zhang, Jure Leskovec
Abstract: GNN-to-MLP distillation aims to utilize knowledge distillation (KD) to learn computationally-efficient multi-layer perceptron (student MLP) on graph data by mimicking the output representations of teacher GNN. Existing methods mainly make the MLP to mimic the GNN predictions over a few class labels. However, the class space may not be expressive enough for covering numerous diverse local graph structures, thus limiting the performance of knowledge transfer from GNN to MLP. To address this issue, we propose to learn a new powerful graph representation space by directly labeling nodes' diverse local structures for GNN-to-MLP distillation. Specifically, we propose a variant of VQ-VAE to learn a structure-aware tokenizer on graph data that can encode each node's local substructure as a discrete code. The discrete codes constitute a codebook as a new graph representation space that is able to identify different local graph structures of nodes with the corresponding code indices. Then, based on the learned codebook, we propose a new distillation target, namely soft code assignments, to directly transfer the structural knowledge of each node from GNN to MLP. The resulting framework VQGraph achieves new state-of-the-art performance on GNN-to-MLP distillation in both transductive and inductive settings across seven graph datasets. We show that VQGraph with better performance infers faster than GNNs by 828x, and also achieves accuracy improvement over GNNs and stand-alone MLPs by 3.90% and 28.05% on average, respectively. Code: https://github.com/YangLing0818/VQGraph.
Authors: Nhat-Tan Bui, Dinh-Hieu Hoang, Minh-Triet Tran, Gianfranco Doretto, Donald Adjeroh, Brijesh Patel, Arabinda Choudhary, Ngan Le
Abstract: Image segmentation remains a pivotal component in medical image analysis, aiding in the extraction of critical information for precise diagnostic practices. With the advent of deep learning, automated image segmentation methods have risen to prominence, showcasing exceptional proficiency in processing medical imagery. Motivated by the Segment Anything Model (SAM)-a foundational model renowned for its remarkable precision and robust generalization capabilities in segmenting 2D natural images-we introduce SAM3D, an innovative adaptation tailored for 3D volumetric medical image analysis. Unlike current SAM-based methods that segment volumetric data by converting the volume into separate 2D slices for individual analysis, our SAM3D model processes the entire 3D volume image in a unified approach. Extensive experiments are conducted on multiple medical image datasets to demonstrate that our network attains competitive results compared with other state-of-the-art methods in 3D medical segmentation tasks while being significantly efficient in terms of parameters. Code and checkpoints are available at https://github.com/UARK-AICV/SAM3D.
Authors: \c{C}a\u{g}lar K\"u\c{c}\"uk, Apostolos Giannakos, Stefan Schneider, Alexander Jann
Abstract: Weather radar data are critical for nowcasting and an integral component of numerical weather prediction models. While weather radar data provide valuable information at high resolution, their ground-based nature limits their availability, which impedes large-scale applications. In contrast, meteorological satellites cover larger domains but with coarser resolution. However, with the rapid advancements in data-driven methodologies and modern sensors aboard geostationary satellites, new opportunities are emerging to bridge the gap between ground- and space-based observations, ultimately leading to more skillful weather prediction with high accuracy. Here, we present a Transformer-based model for nowcasting ground-based radar image sequences using satellite data up to two hours lead time. Trained on a dataset reflecting severe weather conditions, the model predicts radar fields occurring under different weather phenomena and shows robustness against rapidly growing/decaying fields and complex field structures. Model interpretation reveals that the infrared channel centered at 10.3 $\mu m$ (C13) contains skillful information for all weather conditions, while lightning data have the highest relative feature importance in severe weather conditions, particularly in shorter lead times. The model can support precipitation nowcasting across large domains without an explicit need for radar towers, enhance numerical weather prediction and hydrological models, and provide radar proxy for data-scarce regions. Moreover, the open-source framework facilitates progress towards operational data-driven nowcasting.
Authors: Yixiong Chen
Abstract: Image classification is a crucial task in machine learning with widespread practical applications. The existing classical framework for image classification typically utilizes a global pooling operation at the end of the network to reduce computational complexity and mitigate overfitting. However, this operation often results in a significant loss of information, which can affect the performance of classification models. To overcome this limitation, we introduce a novel image classification framework that leverages variational quantum algorithms (VQAs)-hybrid approaches combining quantum and classical computing paradigms within quantum machine learning. The major advantage of our framework is the elimination of the need for the global pooling operation at the end of the network. In this way, our approach preserves more discriminative features and fine-grained details in the images, which enhances classification performance. Additionally, employing VQAs enables our framework to have fewer parameters than the classical framework, even in the absence of global pooling, which makes it more advantageous in preventing overfitting. We apply our method to different state-of-the-art image classification models and demonstrate the superiority of the proposed quantum architecture over its classical counterpart through a series of experiments on public datasets. Our experiments show that the proposed quantum framework achieves up to a 9.21% increase in accuracy and up to a 15.79% improvement in F1 score, compared to the classical framework.
Authors: Haibo Jin, Ruoxi Chen, Andy Zhou, Jinyin Chen, Yang Zhang, Haohan Wang
Abstract: The discovery of "jailbreaks" to bypass safety filters of Large Language Models (LLMs) and harmful responses have encouraged the community to implement safety measures. One major safety measure is to proactively test the LLMs with jailbreaks prior to the release. Therefore, such testing will require a method that can generate jailbreaks massively and efficiently. In this paper, we follow a novel yet intuitive strategy to generate jailbreaks in the style of the human generation. We propose a role-playing system that assigns four different roles to the user LLMs to collaborate on new jailbreaks. Furthermore, we collect existing jailbreaks and split them into different independent characteristics using clustering frequency and semantic patterns sentence by sentence. We organize these characteristics into a knowledge graph, making them more accessible and easier to retrieve. Our system of different roles will leverage this knowledge graph to generate new jailbreaks, which have proved effective in inducing LLMs to generate unethical or guideline-violating responses. In addition, we also pioneer a setting in our system that will automatically follow the government-issued guidelines to generate jailbreaks to test whether LLMs follow the guidelines accordingly. We refer to our system as GUARD (Guideline Upholding through Adaptive Role-play Diagnostics). We have empirically validated the effectiveness of GUARD on three cutting-edge open-sourced LLMs (Vicuna-13B, LongChat-7B, and Llama-2-7B), as well as a widely-utilized commercial LLM (ChatGPT). Moreover, our work extends to the realm of vision language models (MiniGPT-v2 and Gemini Vision Pro), showcasing GUARD's versatility and contributing valuable insights for the development of safer, more reliable LLM-based applications across diverse modalities.
Authors: Jiarun Liu, Hao Yang, Hong-Yu Zhou, Yan Xi, Lequan Yu, Yizhou Yu, Yong Liang, Guangming Shi, Shaoting Zhang, Hairong Zheng, Shanshan Wang
Abstract: Accurate medical image segmentation demands the integration of multi-scale information, spanning from local features to global dependencies. However, it is challenging for existing methods to model long-range global information, where convolutional neural networks (CNNs) are constrained by their local receptive fields, and vision transformers (ViTs) suffer from high quadratic complexity of their attention mechanism. Recently, Mamba-based models have gained great attention for their impressive ability in long sequence modeling. Several studies have demonstrated that these models can outperform popular vision models in various tasks, offering higher accuracy, lower memory consumption, and less computational burden. However, existing Mamba-based models are mostly trained from scratch and do not explore the power of pretraining, which has been proven to be quite effective for data-efficient medical image analysis. This paper introduces a novel Mamba-based model, Swin-UMamba, designed specifically for medical image segmentation tasks, leveraging the advantages of ImageNet-based pretraining. Our experimental results reveal the vital role of ImageNet-based training in enhancing the performance of Mamba-based models. Swin-UMamba demonstrates superior performance with a large margin compared to CNNs, ViTs, and latest Mamba-based models. Notably, on AbdomenMRI, Encoscopy, and Microscopy datasets, Swin-UMamba outperforms its closest counterpart U-Mamba_Enc by an average score of 2.72%.
Authors: Wangjie Zhong, Leimin Tian, Duy Tho Le, Hamid Rezatofighi
Abstract: Social robots often rely on visual perception to understand their users and the environment. Recent advancements in data-driven approaches for computer vision have demonstrated great potentials for applying deep-learning models to enhance a social robot's visual perception. However, the high computational demands of deep-learning methods, as opposed to the more resource-efficient shallow-learning models, bring up important questions regarding their effects on real-world interaction and user experience. It is unclear how will the objective interaction performance and subjective user experience be influenced when a social robot adopts a deep-learning based visual perception model. We employed state-of-the-art human perception and tracking models to improve the visual perception function of the Pepper robot and conducted a controlled lab study and an in-the-wild human-robot interaction study to evaluate this novel perception function for following a specific user with other people present in the scene.
Authors: Yue Yang, Yuqi lin, Hong Liu, Wenqi Shao, Runjian Chen, Hailong Shang, Yu Wang, Yu Qiao, Kaipeng Zhang, Ping Luo
Abstract: Recent text-to-image (T2I) models have had great success, and many benchmarks have been proposed to evaluate their performance and safety. However, they only consider explicit prompts while neglecting implicit prompts (hint at a target without explicitly mentioning it). These prompts may get rid of safety constraints and pose potential threats to the applications of these models. This position paper highlights the current state of T2I models toward implicit prompts. We present a benchmark named ImplicitBench and conduct an investigation on the performance and impacts of implicit prompts with popular T2I models. Specifically, we design and collect more than 2,000 implicit prompts of three aspects: General Symbols, Celebrity Privacy, and Not-Safe-For-Work (NSFW) Issues, and evaluate six well-known T2I models' capabilities under these implicit prompts. Experiment results show that (1) T2I models are able to accurately create various target symbols indicated by implicit prompts; (2) Implicit prompts bring potential risks of privacy leakage for T2I models. (3) Constraints of NSFW in most of the evaluated T2I models can be bypassed with implicit prompts. We call for increased attention to the potential and risks of implicit prompts in the T2I community and further investigation into the capabilities and impacts of implicit prompts, advocating for a balanced approach that harnesses their benefits while mitigating their risks.