Authors: Guojin Cao, Jiaxu Li, Jia He, Ying Min, Yunhao Zhang
Abstract: This technical report presents the implementation details of 2nd winning for CVPR'24 UG2 WeatherProof Dataset Challenge. This challenge aims at semantic segmentation of images degraded by various degrees of weather from all around the world. We addressed this problem by introducing a pre-trained large-scale vision foundation model: InternImage, and trained it using images with different levels of noise. Besides, we did not use additional datasets in the training procedure and utilized dense-CRF as post-processing in the final testing procedure. As a result, we achieved 2nd place in the challenge with 45.1 mIOU and fewer submissions than the other winners.
Authors: Zenglin Shi, Tong Su, Pei Liu, Yunpeng Wu, Le Zhang, Meng Wang
Abstract: This work aims to tackle the all-in-one image restoration task, which seeks to handle multiple types of degradation with a single model. The primary challenge is to extract degradation representations from the input degraded images and use them to guide the model's adaptation to specific degradation types. Recognizing that various degradations affect image content differently across frequency bands, we propose a new all-in-one image restoration approach from a frequency perspective, leveraging advanced vision transformers. Our method consists of two main components: a frequency-aware Degradation prior learning transformer (Dformer) and a degradation-adaptive Restoration transformer (Rformer). The Dformer captures the essential characteristics of various degradations by decomposing inputs into different frequency components. By understanding how degradations affect these frequency components, the Dformer learns robust priors that effectively guide the restoration process. The Rformer then employs a degradation-adaptive self-attention module to selectively focus on the most affected frequency components, guided by the learned degradation representations. Extensive experimental results demonstrate that our approach outperforms the existing methods on four representative restoration tasks, including denoising, deraining, dehazing and deblurring. Additionally, our method offers benefits for handling spatially variant degradations and unseen degradation levels.
Authors: Qingwen Zhang, Yi Yang, Peizheng Li, Olov Andersson, Patric Jensfelt
Abstract: Scene flow estimation predicts the 3D motion at each point in successive LiDAR scans. This detailed, point-level, information can help autonomous vehicles to accurately predict and understand dynamic changes in their surroundings. Current state-of-the-art methods require annotated data to train scene flow networks and the expense of labeling inherently limits their scalability. Self-supervised approaches can overcome the above limitations, yet face two principal challenges that hinder optimal performance: point distribution imbalance and disregard for object-level motion constraints. In this paper, we propose SeFlow, a self-supervised method that integrates efficient dynamic classification into a learning-based scene flow pipeline. We demonstrate that classifying static and dynamic points helps design targeted objective functions for different motion patterns. We also emphasize the importance of internal cluster consistency and correct object point association to refine the scene flow estimation, in particular on object details. Our real-time capable method achieves state-of-the-art performance on the self-supervised scene flow task on Argoverse 2 and Waymo datasets. The code is open-sourced at https://github.com/KTH-RPL/SeFlow along with trained model weights.
Authors: Sebastian A. Cruz Romero, Ivanelyz Rivera de Jesus, Dariana J. Troche Quinones, Wilson Rivera Gallego
Abstract: X-ray image-based disease diagnosis lies in ensuring the precision of identifying afflictions within the sample, a task fraught with challenges stemming from the occurrence of false positives and false negatives. False positives introduce the risk of erroneously identifying non-existent conditions, leading to misdiagnosis and a decline in patient care quality. Conversely, false negatives pose the threat of overlooking genuine abnormalities, potentially causing delays in treatment and interventions, thereby resulting in adverse patient outcomes. The urgency to overcome these challenges compels ongoing efforts to elevate the precision and reliability of X-ray image analysis algorithms within the computational framework. This study introduces modified pre-trained ResNet models tailored for multi-class disease diagnosis of X-ray images, incorporating advanced optimization strategies to reduce the execution runtime of training and inference tasks. The primary objective is to achieve tangible performance improvements through accelerated implementations of PyTorch, CUDA, Mixed- Precision Training, and Learning Rate Scheduler. While outcomes demonstrate substantial improvements in execution runtimes between normal training and CUDA-accelerated training, negligible differences emerge between various training optimization modalities. This research marks a significant advancement in optimizing computational approaches to reduce training execution time for larger models. Additionally, we explore the potential of effective parallel data processing using MPI4Py for the distribution of gradient descent optimization across multiple nodes and leverage multiprocessing to expedite data preprocessing for larger datasets.
Authors: Ahmad AlMughrabi, Umair Haroon, Ricardo Marques, Petia Radeva
Abstract: Accurate food volume estimation is essential for dietary assessment, nutritional tracking, and portion control applications. We present VolETA, a sophisticated methodology for estimating food volume using 3D generative techniques. Our approach creates a scaled 3D mesh of food objects using one- or few-RGBD images. We start by selecting keyframes based on the RGB images and then segmenting the reference object in the RGB images using XMem++. Simultaneously, camera positions are estimated and refined using the PixSfM technique. The segmented food images, reference objects, and camera poses are combined to form a data model suitable for NeuS2. Independent mesh reconstructions for reference and food objects are carried out, with scaling factors determined using MeshLab based on the reference object. Moreover, depth information is used to fine-tune the scaling factors by estimating the potential volume range. The fine-tuned scaling factors are then applied to the cleaned food meshes for accurate volume measurements. Similarly, we enter a segmented RGB image to the One-2-3-45 model for one-shot food volume estimation, resulting in a mesh. We then leverage the obtained scaling factors to the cleaned food mesh for accurate volume measurements. Our experiments show that our method effectively addresses occlusions, varying lighting conditions, and complex food geometries, achieving robust and accurate volume estimations with 10.97% MAPE using the MTF dataset. This innovative approach enhances the precision of volume assessments and significantly contributes to computational nutrition and dietary monitoring advancements.
Authors: Rongzhen Zhao, Vivienne Wang, Juho Kannala, Joni Pajarinen
Abstract: Similar to humans perceiving visual scenes as objects, Object-Centric Learning (OCL) can abstract dense images or videos into sparse object-level features. Transformer-based OCL handles complex textures well due to the decoding guidance of discrete representation, obtained by discretizing noisy features in image or video feature maps using template features from a codebook. However, treating features as minimal units overlooks their composing attributes, thus impeding model generalization; indexing features with natural numbers loses attribute-level commonalities and characteristics, thus diminishing heuristics for model convergence. We propose \textit{Grouped Discrete Representation} (GDR) to address these issues by grouping features into attributes and indexing them with tuple numbers. In extensive experiments across different query initializations, dataset modalities, and model architectures, GDR consistently improves convergence and generalizability. Visualizations show that our method effectively captures attribute-level information in features. The source code will be available upon acceptance.
Authors: Kehinde Ajayi, Leizhen Zhang, Yi He, Jian Wu
Abstract: Quantifying uncertainties for machine learning models is a critical step to reduce human verification effort by detecting predictions with low confidence. This paper proposes a method for uncertainty quantification (UQ) of table structure recognition (TSR). The proposed UQ method is built upon a mixture-of-expert approach termed Test-Time Augmentation (TTA). Our key idea is to enrich and diversify the table representations, to spotlight the cells with high recognition uncertainties. To evaluate the effectiveness, we proposed two heuristics to differentiate highly uncertain cells from normal cells, namely, masking and cell complexity quantification. Masking involves varying the pixel intensity to deem the detection uncertainty. Cell complexity quantification gauges the uncertainty of each cell by its topological relation with neighboring cells. The evaluation results based on standard benchmark datasets demonstrate that the proposed method is effective in quantifying uncertainty in TSR models. To our best knowledge, this study is the first of its kind to enable UQ in TSR tasks. Our code and data are available at: https://github.com/lamps-lab/UQTTA.git.
Authors: Yujin Ham, Mateusz Michalkiewicz, Guha Balakrishnan
Abstract: 3D building reconstruction from imaging data is an important task for many applications ranging from urban planning to reconnaissance. Modern Novel View synthesis (NVS) methods like NeRF and Gaussian Splatting offer powerful techniques for developing 3D models from natural 2D imagery in an unsupervised fashion. These algorithms generally require input training views surrounding the scene of interest, which, in the case of large buildings, is typically not available across all camera elevations. In particular, the most readily available camera viewpoints at scale across most buildings are at near-ground (e.g., with mobile phones) and aerial (drones) elevations. However, due to the significant difference in viewpoint between drone and ground image sets, camera registration - a necessary step for NVS algorithms - fails. In this work we propose a method, DRAGON, that can take drone and ground building imagery as input and produce a 3D NVS model. The key insight of DRAGON is that intermediate elevation imagery may be extrapolated by an NVS algorithm itself in an iterative procedure with perceptual regularization, thereby bridging the visual feature gap between the two elevations and enabling registration. We compiled a semi-synthetic dataset of 9 large building scenes using Google Earth Studio, and quantitatively and qualitatively demonstrate that DRAGON can generate compelling renderings on this dataset compared to baseline strategies.
Authors: Francis Williams, Jiahui Huang, Jonathan Swartz, Gergely Kl\'ar, Vijay Thakkar, Matthew Cong, Xuanchi Ren, Ruilong Li, Clement Fuji-Tsang, Sanja Fidler, Eftychios Sifakis, Ken Museth
Abstract: We present fVDB, a novel GPU-optimized framework for deep learning on large-scale 3D data. fVDB provides a complete set of differentiable primitives to build deep learning architectures for common tasks in 3D learning such as convolution, pooling, attention, ray-tracing, meshing, etc. fVDB simultaneously provides a much larger feature set (primitives and operators) than established frameworks with no loss in efficiency: our operators match or exceed the performance of other frameworks with narrower scope. Furthermore, fVDB can process datasets with much larger footprint and spatial resolution than prior works, while providing a competitive memory footprint on small inputs. To achieve this combination of versatility and performance, fVDB relies on a single novel VDB index grid acceleration structure paired with several key innovations including GPU accelerated sparse grid construction, convolution using tensorcores, fast ray tracing kernels using a Hierarchical Digital Differential Analyzer algorithm (HDDA), and jagged tensors. Our framework is fully integrated with PyTorch enabling interoperability with existing pipelines, and we demonstrate its effectiveness on a number of representative tasks such as large-scale point-cloud segmentation, high resolution 3D generative modeling, unbounded scale Neural Radiance Fields, and large-scale point cloud reconstruction.
Authors: Ali Borji
Abstract: The primary aim of this manuscript is to underscore a significant limitation in current deep learning models, particularly vision models. Unlike human vision, which efficiently selects only the essential visual areas for further processing, leading to high speed and low energy consumption, deep vision models process the entire image. In this work, we examine this issue from a broader perspective and propose a solution that could pave the way for the next generation of more efficient vision models. Basically, convolution and pooling operations are selectively applied to altered regions, with a change map sent to subsequent layers. This map indicates which computations need to be repeated. The code is available at https://github.com/aliborji/spatial_attention.
Authors: Ali Karaoglu
Abstract: In contemporary imaging systems, achieving optimal auto-focus (AF) performance hinges on precise lens positioning. Extensive research has delved into refining algorithms for determining the ideal lens position across passive, active, and hybrid autofocus systems. This paper explores the mathematical intricacies and practical considerations essential for optimizing lens positions during focus searches, addressing overarching challenges encountered in AF systems, such as balancing speed and accuracy. Moreover, the lens position calculations proposed herein can be applied to various focus algorithms, including focus bracketing. The proposed method offers adaptability and scalability, rendering it suitable for integration into a wide array of camera systems, ranging from smartphones and DSLRs to microscopes and industrial imaging devices.
Authors: Jiayi Wang, Kevin Alexander Laube, Yumeng Li, Jan Hendrik Metzen, Shin-I Cheng, Julio Borges, Anna Khoreva
Abstract: Recent work has shown great progress in integrating spatial conditioning to control large, pre-trained text-to-image diffusion models. Despite these advances, existing methods describe the spatial image content using hand-crafted conditioning inputs, which are either semantically ambiguous (e.g., edges) or require expensive manual annotations (e.g., semantic segmentation). To address these limitations, we propose a new label-free way of conditioning diffusion models to enable fine-grained spatial control. We introduce the concept of neural semantic image synthesis, which uses neural layouts extracted from pre-trained foundation models as conditioning. Neural layouts are advantageous as they provide rich descriptions of the desired image, containing both semantics and detailed geometry of the scene. We experimentally show that images synthesized via neural semantic image synthesis achieve similar or superior pixel-level alignment of semantic classes compared to those created using expensive semantic label maps. At the same time, they capture better semantics, instance separation, and object orientation than other label-free conditioning options, such as edges or depth. Moreover, we show that images generated by neural layout conditioning can effectively augment real data for training various perception tasks.
Authors: Alejandro Lozano, Jeffrey Nirschl, James Burgess, Sanket Rajan Gupte, Yuhui Zhang, Alyssa Unell, Serena Yeung-Levy
Abstract: Recent advances in microscopy have enabled the rapid generation of terabytes of image data in cell biology and biomedical research. Vision-language models (VLMs) offer a promising solution for large-scale biological image analysis, enhancing researchers' efficiency, identifying new image biomarkers, and accelerating hypothesis generation and scientific discovery. However, there is a lack of standardized, diverse, and large-scale vision-language benchmarks to evaluate VLMs' perception and cognition capabilities in biological image understanding. To address this gap, we introduce {\mu}-Bench, an expert-curated benchmark encompassing 22 biomedical tasks across various scientific disciplines (biology, pathology), microscopy modalities (electron, fluorescence, light), scales (subcellular, cellular, tissue), and organisms in both normal and abnormal states. We evaluate state-of-the-art biomedical, pathology, and general VLMs on {\mu}-Bench and find that: i) current models struggle on all categories, even for basic tasks such as distinguishing microscopy modalities; ii) current specialist models fine-tuned on biomedical data often perform worse than generalist models; iii) fine-tuning in specific microscopy domains can cause catastrophic forgetting, eroding prior biomedical knowledge encoded in their base model. iv) weight interpolation between fine-tuned and pre-trained models offers one solution to forgetting and improves general performance across biomedical tasks. We release {\mu}-Bench under a permissive license to accelerate the research and development of microscopy foundation models.
Authors: Aneeshan Sain, Pinaki Nath Chowdhury, Subhadeep Koley, Ayan Kumar Bhunia, Yi-Zhe Song
Abstract: In this paper, we delve into the intricate dynamics of Fine-Grained Sketch-Based Image Retrieval (FG-SBIR) by addressing a critical yet overlooked aspect -- the choice of viewpoint during sketch creation. Unlike photo systems that seamlessly handle diverse views through extensive datasets, sketch systems, with limited data collected from fixed perspectives, face challenges. Our pilot study, employing a pre-trained FG-SBIR model, highlights the system's struggle when query-sketches differ in viewpoint from target instances. Interestingly, a questionnaire however shows users desire autonomy, with a significant percentage favouring view-specific retrieval. To reconcile this, we advocate for a view-aware system, seamlessly accommodating both view-agnostic and view-specific tasks. Overcoming dataset limitations, our first contribution leverages multi-view 2D projections of 3D objects, instilling cross-modal view awareness. The second contribution introduces a customisable cross-modal feature through disentanglement, allowing effortless mode switching. Extensive experiments on standard datasets validate the effectiveness of our method.
Authors: Jinjing Zhu, Yucheng Chen, Lin Wang
Abstract: Unsupervised domain adaption (UDA) has emerged as a popular solution to tackle the divergence between the labeled source and unlabeled target domains. Recently, some research efforts have been made to leverage large vision-language models, such as CLIP, and then fine-tune or learn prompts from them for addressing the challenging UDA task. In this work, we shift the gear to a new direction by directly leveraging CLIP to measure the domain divergence and propose a novel language-guided approach for UDA, dubbed as CLIP-Div. Our key idea is to harness CLIP to 1) measure the domain divergence via the acquired domain-agnostic distribution and 2) calibrate the target pseudo labels with language guidance, to effectively reduce the domain gap and improve the UDA model's generalization capability. Specifically, our major technical contribution lies in the proposed two novel language-guided domain divergence measurement losses: absolute divergence and relative divergence. These loss terms furnish precise guidelines for aligning the distributions of the source and target domains with the domain-agnostic distribution derived from CLIP. Additionally, we propose a language-guided pseudo-labeling strategy for calibrating the target pseudo labels. Buttressed by it, we show that a further implementation for self-training can enhance the UDA model's generalization capability on the target domain. CLIP-Div surpasses state-of-the-art CNN-based methods by a substantial margin, achieving a performance boost of +10.3% on Office-Home, +1.5% on Office-31, +0.2% on VisDA-2017, and +24.3% on DomainNet, respectively.
Authors: Pratyush Tripathy, Kathy Baylis, Kyle Wu, Jyles Watson, Ruizhe Jiang
Abstract: Accurate mapping of agricultural field boundaries is crucial for enhancing outcomes like precision agriculture, crop monitoring, and yield estimation. However, extracting these boundaries from satellite images is challenging, especially for smallholder farms and data-scarce environments. This study explores the Segment Anything Model (SAM) to delineate agricultural field boundaries in Bihar, India, using 2-meter resolution SkySat imagery without additional training. We evaluate SAM's performance across three model checkpoints, various input sizes, multi-date satellite images, and edge-enhanced imagery. Our results show that SAM correctly identifies about 58% of field boundaries, comparable to other approaches requiring extensive training data. Using different input image sizes improves accuracy, with the most significant improvement observed when using multi-date satellite images. This work establishes proof of concept for using SAM and maximizing its potential in agricultural field boundary mapping. Our work highlights SAM's potential in delineating agriculture field boundary in training-data scarce settings to enable a wide range of agriculture related analysis.
Authors: Sanjoy Chowdhury, Sayan Nag, Subhrajyoti Dasgupta, Jun Chen, Mohamed Elhoseiny, Ruohan Gao, Dinesh Manocha
Abstract: Leveraging Large Language Models' remarkable proficiency in text-based tasks, recent works on Multi-modal LLMs (MLLMs) extend them to other modalities like vision and audio. However, the progress in these directions has been mostly focused on tasks that only require a coarse-grained understanding of the audio-visual semantics. We present Meerkat, an audio-visual LLM equipped with a fine-grained understanding of image and audio both spatially and temporally. With a new modality alignment module based on optimal transport and a cross-attention module that enforces audio-visual consistency, Meerkat can tackle challenging tasks such as audio referred image grounding, image guided audio temporal localization, and audio-visual fact-checking. Moreover, we carefully curate a large dataset AVFIT that comprises 3M instruction tuning samples collected from open-source datasets, and introduce MeerkatBench that unifies five challenging audio-visual tasks. We achieve state-of-the-art performance on all these downstream tasks with a relative improvement of up to 37.12%.
Authors: Shiquan Shen, Zhizhong Wu, Pan Zhang
Abstract: With the development of deep learning technology, the detection and classification of distracted driving behaviour requires higher accuracy. Existing deep learning-based methods are computationally intensive and parameter redundant, limiting the efficiency and accuracy in practical applications. To solve this problem, this study proposes an improved YOLOv8 detection method based on the original YOLOv8 model by integrating the BoTNet module, GAM attention mechanism and EIoU loss function. By optimising the feature extraction and multi-scale feature fusion strategies, the training and inference processes are simplified, and the detection accuracy and efficiency are significantly improved. Experimental results show that the improved model performs well in both detection speed and accuracy, with an accuracy rate of 99.4%, and the model is smaller and easy to deploy, which is able to identify and classify distracted driving behaviours in real time, provide timely warnings, and enhance driving safety.
Authors: Yunxiang Zhang, Alexandr Kuznetsov, Akshay Jindal, Kenneth Chen, Anton Sochenov, Anton Kaplanyan, Qi Sun
Abstract: Neural image representations have recently emerged as a promising technique for storing, streaming, and rendering visual data. Coupled with learning-based workflows, these novel representations have demonstrated remarkable visual fidelity and memory efficiency. However, existing neural image representations often rely on explicit uniform data structures without content adaptivity or computation-intensive implicit models, limiting their adoption in real-time graphics applications. Inspired by recent advances in radiance field rendering, we propose Image-GS, a content-adaptive image representation. Using anisotropic 2D Gaussians as the basis, Image-GS shows high memory efficiency, supports fast random access, and offers a natural level of detail stack. Leveraging a tailored differentiable renderer, Image-GS fits a target image by adaptively allocating and progressively optimizing a set of 2D Gaussians. The generalizable efficiency and fidelity of Image-GS are validated against several recent neural image representations and industry-standard texture compressors on a diverse set of images. Notably, its memory and computation requirements solely depend on and linearly scale with the number of 2D Gaussians, providing flexible controls over the trade-off between visual fidelity and run-time efficiency. We hope this research offers insights for developing new applications that require adaptive quality and resource control, such as machine perception, asset streaming, and content generation.
Authors: Kunyu Peng, Jia Fu, Kailun Yang, Di Wen, Yufan Chen, Ruiping Liu, Junwei Zheng, Jiaming Zhang, M. Saquib Sarfraz, Rainer Stiefelhagen, Alina Roitberg
Abstract: We introduce a new task called Referring Atomic Video Action Recognition (RAVAR), aimed at identifying atomic actions of a particular person based on a textual description and the video data of this person. This task differs from traditional action recognition and localization, where predictions are delivered for all present individuals. In contrast, we focus on recognizing the correct atomic action of a specific individual, guided by text. To explore this task, we present the RefAVA dataset, containing 36,630 instances with manually annotated textual descriptions of the individuals. To establish a strong initial benchmark, we implement and validate baselines from various domains, e.g., atomic action localization, video question answering, and text-video retrieval. Since these existing methods underperform on RAVAR, we introduce RefAtomNet -- a novel cross-stream attention-driven method specialized for the unique challenges of RAVAR: the need to interpret a textual referring expression for the targeted individual, utilize this reference to guide the spatial localization and harvest the prediction of the atomic actions for the referring person. The key ingredients are: (1) a multi-stream architecture that connects video, text, and a new location-semantic stream, and (2) cross-stream agent attention fusion and agent token fusion which amplify the most relevant information across these streams and consistently surpasses standard attention-based fusion on RAVAR. Extensive experiments demonstrate the effectiveness of RefAtomNet and its building blocks for recognizing the action of the described individual. The dataset and code will be made publicly available at https://github.com/KPeng9510/RAVAR.
Authors: Xu Zheng, Ling Wang, Kanghao Chen, Yuanhuiyi Lyu, Jiazhou Zhou, Lin Wang
Abstract: Recently, electroencephalography (EEG) signals have been actively incorporated to decode brain activity to visual or textual stimuli and achieve object recognition in multi-modal AI. Accordingly, endeavors have been focused on building EEG-based datasets from visual or textual single-modal stimuli. However, these datasets offer limited EEG epochs per category, and the complex semantics of stimuli presented to participants compromise their quality and fidelity in capturing precise brain activity. The study in neuroscience unveils that the relationship between visual and textual stimulus in EEG recordings provides valuable insights into the brain's ability to process and integrate multi-modal information simultaneously. Inspired by this, we propose a novel large-scale multi-modal dataset, named EIT-1M, with over 1 million EEG-image-text pairs. Our dataset is superior in its capacity of reflecting brain activities in simultaneously processing multi-modal information. To achieve this, we collected data pairs while participants viewed alternating sequences of visual-textual stimuli from 60K natural images and category-specific texts. Common semantic categories are also included to elicit better reactions from participants' brains. Meanwhile, response-based stimulus timing and repetition across blocks and sessions are included to ensure data diversity. To verify the effectiveness of EIT-1M, we provide an in-depth analysis of EEG data captured from multi-modal stimuli across different categories and participants, along with data quality scores for transparency. We demonstrate its validity on two tasks: 1) EEG recognition from visual or textual stimuli or both and 2) EEG-to-visual generation.
Authors: Zixing Li, Chao Yan, Zhen Lan, Dengqing Tang, Xiaojia Xiang, Han Zhou, Jun Lai
Abstract: Advanced cognition can be extracted from the human brain using brain-computer interfaces. Integrating these interfaces with computer vision techniques, which possess efficient feature extraction capabilities, can achieve more robust and accurate detection of dim targets in aerial images. However, existing target detection methods primarily concentrate on homogeneous data, lacking efficient and versatile processing capabilities for heterogeneous multimodal data. In this paper, we first build a brain-eye-computer based object detection system for aerial images under few-shot conditions. This system detects suspicious targets using region proposal networks, evokes the event-related potential (ERP) signal in electroencephalogram (EEG) through the eye-tracking-based slow serial visual presentation (ESSVP) paradigm, and constructs the EEG-image data pairs with eye movement data. Then, an adaptive modality balanced online knowledge distillation (AMBOKD) method is proposed to recognize dim objects with the EEG-image data. AMBOKD fuses EEG and image features using a multi-head attention module, establishing a new modality with comprehensive features. To enhance the performance and robust capability of the fusion modality, simultaneous training and mutual learning between modalities are enabled by end-to-end online knowledge distillation. During the learning process, an adaptive modality balancing module is proposed to ensure multimodal equilibrium by dynamically adjusting the weights of the importance and the training gradients across various modalities. The effectiveness and superiority of our method are demonstrated by comparing it with existing state-of-the-art methods. Additionally, experiments conducted on public datasets and system validations in real-world scenarios demonstrate the reliability and practicality of the proposed system and the designed method.
Authors: Jiawei Zhan, Jinxiang Lai, Bin-Bin Gao, Jun Liu, Xiaochen Chen, Chengjie Wang
Abstract: Anomaly detection, the technique of identifying abnormal samples using only normal samples, has attracted widespread interest in industry. Existing one-model-per-category methods often struggle with limited generalization capabilities due to their focus on a single category, and can fail when encountering variations in product. Recent feature reconstruction methods, as representatives in one-model-all-categories schemes, face challenges including reconstructing anomalous samples and blurry reconstructions. In this paper, we creatively combine a diffusion model and a transformer for multi-class anomaly detection. This approach leverages diffusion to obtain high-frequency information for refinement, greatly alleviating the blurry reconstruction problem while maintaining the sampling efficiency of the reverse diffusion process. The task is transformed into image inpainting to disconnect the input-output correlation, thereby mitigating the "identical shortcuts" problem and avoiding the model from reconstructing anomalous samples. Besides, we introduce category-awareness using dual conditioning to ensure the accuracy of prediction and reconstruction in the reverse diffusion process, preventing excessive deviation from the target category, thus effectively enabling multi-class anomaly detection. Futhermore, Spatio-temporal fusion is also employed to fuse heatmaps predicted at different timesteps and scales, enhancing the performance of multi-class anomaly detection. Extensive experiments on benchmark datasets demonstrate the superior performance and exceptional multi-class anomaly detection capabilities of our proposed method compared to others.
Authors: Hailiang Zhang, Dian Chao, Zhihao Guan, Yang Yang
Abstract: In this paper, we introduce a grounded video question-answering solution. Our research reveals that the fixed official baseline method for video question answering involves two main steps: visual grounding and object tracking. However, a significant challenge emerges during the initial step, where selected frames may lack clearly identifiable target objects. Furthermore, single images cannot address questions like "Track the container from which the person pours the first time." To tackle this issue, we propose an alternative two-stage approach:(1) First, we leverage the VALOR model to answer questions based on video information.(2) concatenate the answered questions with their respective answers. Finally, we employ TubeDETR to generate bounding boxes for the targets.
Authors: Huanzhang Dou, Ruixiang Li, Wei Su, Xi Li
Abstract: In text-to-video (T2V) generation, significant attention has been directed toward its development, yet unifying discrete and continuous grounding conditions in T2V generation remains under-explored. This paper proposes a Grounded text-to-Video generation framework, termed GVDIFF. First, we inject the grounding condition into the self-attention through an uncertainty-based representation to explicitly guide the focus of the network. Second, we introduce a spatial-temporal grounding layer that connects the grounding condition with target objects and enables the model with the grounded generation capacity in the spatial-temporal domain. Third, our dynamic gate network adaptively skips the redundant grounding process to selectively extract grounding information and semantics while improving efficiency. We extensively evaluate the grounded generation capacity of GVDIFF and demonstrate its versatility in applications, including long-range video generation, sequential prompts, and object-specific editing.
Authors: Zijian Ying, Qianmu Li, Tao Wang, Zhichao Lian, Shunmei Meng, Xuyun Zhang
Abstract: Various methods try to enhance adversarial transferability by improving the generalization from different perspectives. In this paper, we rethink the optimization process and propose a novel sequence optimization concept, which is named Looking From the Future (LFF). LFF makes use of the original optimization process to refine the very first local optimization choice. Adapting the LFF concept to the adversarial attack task, we further propose an LFF attack as well as an MLFF attack with better generalization ability. Furthermore, guiding with the LFF concept, we propose an $LLF^{\mathcal{N}}$ attack which entends the LFF attack to a multi-order attack, further enhancing the transfer attack ability. All our proposed methods can be directly applied to the iteration-based attack methods. We evaluate our proposed method on the ImageNet1k dataset by applying several SOTA adversarial attack methods under four kinds of tasks. Experimental results show that our proposed method can greatly enhance the attack transferability. Ablation experiments are also applied to verify the effectiveness of each component. The source code will be released after this paper is accepted.
Authors: Wenlong Liu, Tianyu Yang, Qizhi Yu, Lei Zhang
Abstract: SymPoint is an initial attempt that utilizes point set representation to solve the panoptic symbol spotting task on CAD drawing. Despite its considerable success, it overlooks graphical layer information and suffers from prohibitively slow training convergence. To tackle this issue, we introduce SymPoint-V2, a robust and efficient solution featuring novel, streamlined designs that overcome these limitations. In particular, we first propose a Layer Feature-Enhanced module (LFE) to encode the graphical layer information into the primitive feature, which significantly boosts the performance. We also design a Position-Guided Training (PGT) method to make it easier to learn, which accelerates the convergence of the model in the early stages and further promotes performance. Extensive experiments show that our model achieves better performance and faster convergence than its predecessor SymPoint on the public benchmark. Our code and trained models are available at https://github.com/nicehuster/SymPointV2.
Authors: Yuzheng Wang, Zhaoyu Chen, Dingkang Yang, Yunquan Sun, Lizhe Qi
Abstract: Novel Class Discovery (NCD) aims to discover unknown and novel classes in an unlabeled set by leveraging knowledge already learned about known classes. Existing works focus on instance-level or class-level knowledge representation and build a shared representation space to achieve performance improvements. However, a long-neglected issue is the potential imbalanced number of samples from known and novel classes, pushing the model towards dominant classes. Therefore, these methods suffer from a challenging trade-off between reviewing known classes and discovering novel classes. Based on this observation, we propose a Self-Cooperation Knowledge Distillation (SCKD) method to utilize each training sample (whether known or novel, labeled or unlabeled) for both review and discovery. Specifically, the model's feature representations of known and novel classes are used to construct two disjoint representation spaces. Through spatial mutual information, we design a self-cooperation learning to encourage model learning from the two feature representation spaces from itself. Extensive experiments on six datasets demonstrate that our method can achieve significant performance improvements, achieving state-of-the-art performance.
Authors: Krithika Iyer, Shireen Y. Elhabian
Abstract: The study of physiology demonstrates that the form (shape)of anatomical structures dictates their functions, and analyzing the form of anatomies plays a crucial role in clinical research. Statistical shape modeling (SSM) is a widely used tool for quantitative analysis of forms of anatomies, aiding in characterizing and identifying differences within a population of subjects. Despite its utility, the conventional SSM construction pipeline is often complex and time-consuming. Additionally, reliance on linearity assumptions further limits the model from capturing clinically relevant variations. Recent advancements in deep learning solutions enable the direct inference of SSM from unsegmented medical images, streamlining the process and improving accessibility. However, the new methods of SSM from images do not adequately account for situations where the imaging data quality is poor or where only sparse information is available. Moreover, quantifying aleatoric uncertainty, which represents inherent data variability, is crucial in deploying deep learning for clinical tasks to ensure reliable model predictions and robust decision-making, especially in challenging imaging conditions. Therefore, we propose SPI-CorrNet, a unified model that predicts 3D correspondences from sparse imaging data. It leverages a teacher network to regularize feature learning and quantifies data-dependent aleatoric uncertainty by adapting the network to predict intrinsic input variances. Experiments on the LGE MRI left atrium dataset and Abdomen CT-1K liver datasets demonstrate that our technique enhances the accuracy and robustness of sparse image-driven SSM.
Authors: Zhaoshuai Qi, Yifeng Hao, Rui Hu, Wenyou Chang, Jiaqi Yang, Yanning Zhang
Abstract: Structured light-based method with a camera-projector pair (CPP) plays a vital role in indoor 3D reconstruction, especially for scenes with weak textures. Previous methods usually assume known intrinsics, which are pre-calibrated from known objects, or self-calibrated from multi-view observations. It is still challenging to reliably recover CPP intrinsics from only two views without any known objects. In this paper, we provide a simple yet reliable solution. We demonstrate that, for the first time, sufficient constraints on CPP intrinsics can be derived from an unknown cuboid corner (C2), e.g. a room's corner, which is a common structure in indoor scenes. In addition, with only known camera principal point, the complex multi-variable estimation of all CPP intrinsics can be simplified to a simple univariable optimization problem, leading to reliable calibration and thus direct 3D reconstruction with unknown CPP. Extensive results have demonstrated the superiority of the proposed method over both traditional and learning-based counterparts. Furthermore, the proposed method also demonstrates impressive potential to solve similar tasks without active lighting, such as sparse-view structure from motion.
Authors: Shuo Li, Yubo Cui, Zhiheng Li, Zheng Fang
Abstract: 3D single object tracking (SOT) is a crucial task in fields of mobile robotics and autonomous driving. Traditional motion-based approaches achieve target tracking by estimating the relative movement of target between two consecutive frames. However, they usually overlook local motion information of the target and fail to exploit historical frame information effectively. To overcome the above limitations, we propose a point-level flow method with multi-frame information for 3D SOT task, called FlowTrack. Specifically, by estimating the flow for each point in the target, our method could capture the local motion details of target, thereby improving the tracking performance. At the same time, to handle scenes with sparse points, we present a learnable target feature as the bridge to efficiently integrate target information from past frames. Moreover, we design a novel Instance Flow Head to transform dense point-level flow into instance-level motion, effectively aggregating local motion information to obtain global target motion. Finally, our method achieves competitive performance with improvements of 5.9% on the KITTI dataset and 2.9% on NuScenes. The code will be made publicly available soon.
Authors: Cong Cao, Huanjing Yue, Xin Liu, Jingyu Yang
Abstract: Diffusion-based zero-shot image restoration and enhancement models have achieved great success in various image restoration and enhancement tasks without training. However, directly applying them to video restoration and enhancement results in severe temporal flickering artifacts. In this paper, we propose the first framework for zero-shot video restoration and enhancement based on a pre-trained image diffusion model. By replacing the self-attention layer with the proposed cross-previous-frame attention layer, the pre-trained image diffusion model can take advantage of the temporal correlation between neighboring frames. We further propose temporal consistency guidance, spatial-temporal noise sharing, and an early stopping sampling strategy for better temporally consistent sampling. Our method is a plug-and-play module that can be inserted into any diffusion-based zero-shot image restoration or enhancement methods to further improve their performance. Experimental results demonstrate the superiority of our proposed method in producing temporally consistent videos with better fidelity.
Authors: Shi Tang, Guiming Luo, Xinchen Ye, Zhiyi Xia
Abstract: Generalizing to novel classes unseen during training is a key challenge of few-shot classification. Recent metric-based methods try to address this by local representations. However, they are unable to take full advantage of them due to (i) improper supervision for pretraining the feature extractor, and (ii) lack of adaptability in the metric for handling various possible compositions of local feature sets. In this work, we unleash the power of local representations in improving novel-class generalization. For the feature extractor, we design a novel pretraining paradigm that learns randomly cropped patches by soft labels. It utilizes the class-level diversity of patches while diminishing the impact of their semantic misalignments to hard labels. To align network output with soft labels, we also propose a UniCon KL-Divergence that emphasizes the equal contribution of each base class in describing "non-base" patches. For the metric, we formulate measuring local feature sets as an entropy-regularized optimal transport problem to introduce the ability to handle sets consisting of homogeneous elements. Furthermore, we design a Modulate Module to endow the metric with the necessary adaptability. Our method achieves new state-of-the-art performance on three popular benchmarks. Moreover, it exceeds state-of-the-art transductive and cross-modal methods in the fine-grained scenario.
Authors: Zhiyu Pan, Kewei Wang, Yizheng Wu, Liwen Xiao, Jiahao Cui, Zhicheng Wang, Zhiguo Cao
Abstract: Automatic image cropping models predict reframing boxes to enhance image aesthetics. Yet, the scarcity of labeled data hinders the progress of this task. To overcome this limitation, we explore the possibility of utilizing both labeled and unlabeled data together to expand the scale of training data for image cropping models. This idea can be implemented in a pseudo-labeling way: producing pseudo labels for unlabeled data by a teacher model and training a student model with these pseudo labels. However, the student may learn from teacher's mistakes. To address this issue, we propose the multi-policy viewfinder network (MPV-Net) that offers diverse refining policies to rectify the mistakes in original pseudo labels from the teacher. The most reliable policy is selected to generate trusted pseudo labels. The reliability of policies is evaluated via the robustness against box jittering. The efficacy of our method can be evaluated by the improvement compared to the supervised baseline which only uses labeled data. Notably, our MPV-Net outperforms off-the-shelf pseudo-labeling methods, yielding the most substantial improvement over the supervised baseline. Furthermore, our approach achieves state-of-the-art results on both the FCDB and FLMS datasets, signifying the superiority of our approach.
Authors: Long Hoang Dang, Thao Minh Le, Vuong Le, Tu Minh Phuong, Truyen Tran
Abstract: Large vision-language models (LVLMs) offer a novel capability for performing in-context learning (ICL) in Visual QA. When prompted with a few demonstrations of image-question-answer triplets, LVLMs have demonstrated the ability to discern underlying patterns and transfer this latent knowledge to answer new questions about unseen images without the need for expensive supervised fine-tuning. However, designing effective vision-language prompts, especially for compositional questions, remains poorly understood. Adapting language-only ICL techniques may not necessarily work because we need to bridge the visual-linguistic semantic gap: Symbolic concepts must be grounded in visual content, which does not share the syntactic linguistic structures. This paper introduces SADL, a new visual-linguistic prompting framework for the task. SADL revolves around three key components: SAmpling, Deliberation, and Pseudo-Labeling of image-question pairs. Given an image-question query, we sample image-question pairs from the training data that are in semantic proximity to the query. To address the compositional nature of questions, the deliberation step decomposes complex questions into a sequence of subquestions. Finally, the sequence is progressively annotated one subquestion at a time to generate a sequence of pseudo-labels. We investigate the behaviors of SADL under OpenFlamingo on large-scale Visual QA datasets, namely GQA, GQA-OOD, CLEVR, and CRIC. The evaluation demonstrates the critical roles of sampling in the neighborhood of the image, the decomposition of complex questions, and the accurate pairing of the subquestions and labels. These findings do not always align with those found in language-only ICL, suggesting fresh insights in vision-language settings.
Authors: Long Hoang Dang, Duy-Hung Nguyen, Thai Quang Le, Thinh Truong Nguyen, Clark Mei, Vu Hoang
Abstract: This paper presents a novel system, named AHMsys, designed to automate the process of generating 3D Heating, Ventilation, and Air Conditioning (HVAC) models from 2D Computer-Aided Design (CAD) drawings, a key component of Building Information Modeling (BIM). By automatically preprocessing and extracting essential HVAC object information then creating detailed 3D models, our proposed AHMsys significantly reduced the 20 percent work schedule of the BIM process in Akila. This advancement highlights the essential impact of integrating AI technologies in managing the lifecycle of a digital representation of the building.
Authors: Marani Badr-Eddine, Hanini Mohamed, Malayarukil Nihitha, Christodoulidis Stergios, Vakalopoulou Maria, Ferrante Enzo
Abstract: The proliferation of machine learning models in critical decision making processes has underscored the need for bias discovery and mitigation strategies. Identifying the reasons behind a biased system is not straightforward, since in many occasions they are associated with hidden spurious correlations which are not easy to spot. Standard approaches rely on bias audits performed by analyzing model performance in pre-defined subgroups of data samples, usually characterized by common attributes like gender or ethnicity when it comes to people, or other specific attributes defining semantically coherent groups of images. However, it is not always possible to know a-priori the specific attributes defining the failure modes of visual recognition systems. Recent approaches propose to discover these groups by leveraging large vision language models, which enable the extraction of cross-modal embeddings and the generation of textual descriptions to characterize the subgroups where a certain model is underperforming. In this work, we argue that incorporating visual explanations (e.g. heatmaps generated via GradCAM or other approaches) can boost the performance of such bias discovery and mitigation frameworks. To this end, we introduce Visually Grounded Bias Discovery and Mitigation (ViG-Bias), a simple yet effective technique which can be integrated to a variety of existing frameworks to improve both, discovery and mitigation performance. Our comprehensive evaluation shows that incorporating visual explanations enhances existing techniques like DOMINO, FACTS and Bias-to-Text, across several challenging datasets, including CelebA, Waterbirds, and NICO++.
Authors: Khanh-Binh Nguyen, Chae Jung Park
Abstract: The primary aim of Audio-Visual Segmentation (AVS) is to precisely identify and locate auditory elements within visual scenes by accurately predicting segmentation masks at the pixel level. Achieving this involves comprehensively considering data and model aspects to address this task effectively. This study presents a lightweight approach, SAVE, which efficiently adapts the pre-trained segment anything model (SAM) to the AVS task. By incorporating an image encoder adapter into the transformer blocks to better capture the distinct dataset information and proposing a residual audio encoder adapter to encode the audio features as a sparse prompt, our proposed model achieves effective audio-visual fusion and interaction during the encoding stage. Our proposed method accelerates the training and inference speed by reducing the input resolution from 1024 to 256 pixels while achieving higher performance compared with the previous SOTA. Extensive experimentation validates our approach, demonstrating that our proposed model outperforms other SOTA methods significantly. Moreover, leveraging the pre-trained model on synthetic data enhances performance on real AVSBench data, achieving 84.59 mIoU on the S4 (V1S) subset and 70.28 mIoU on the MS3 (V1M) set with only 256 pixels for input images. This increases up to 86.16 mIoU on the S4 (V1S) and 70.83 mIoU on the MS3 (V1M) with inputs of 1024 pixels.
Authors: Chengchao Shen, Jianzhong Chen, Jianxin Wang
Abstract: The existing contrastive learning methods mainly focus on single-grained representation learning, e.g., part-level, object-level or scene-level ones, thus inevitably neglecting the transferability of representations on other granularity levels. In this paper, we aim to learn multi-grained representations, which can effectively describe the image on various granularity levels, thus improving generalization on extensive downstream tasks. To this end, we propose a novel Multi-Grained Contrast method (MGC) for unsupervised representation learning. Specifically, we construct delicate multi-grained correspondences between positive views and then conduct multi-grained contrast by the correspondences to learn more general unsupervised representations. Without pretrained on large-scale dataset, our method significantly outperforms the existing state-of-the-art methods on extensive downstream tasks, including object detection, instance segmentation, scene parsing, semantic segmentation and keypoint detection. Moreover, experimental results support the data-efficient property and excellent representation transferability of our method. The source code and trained weights are available at \url{https://github.com/visresearch/mgc}.
Authors: Chaofan Luo, Donglin Di, Yongjia Ma, Zhou Xue, Chen Wei, Xun Yang, Yebin Liu
Abstract: Despite significant strides in the field of 3D scene editing, current methods encounter substantial challenge, particularly in preserving 3D consistency in multi-view editing process. To tackle this challenge, we propose a progressive 3D editing strategy that ensures multi-view consistency via a Trajectory-Anchored Scheme (TAS) with a dual-branch editing mechanism. Specifically, TAS facilitates a tightly coupled iterative process between 2D view editing and 3D updating, preventing error accumulation yielded from text-to-image process. Additionally, we explore the relationship between optimization-based methods and reconstruction-based methods, offering a unified perspective for selecting superior design choice, supporting the rationale behind the designed TAS. We further present a tuning-free View-Consistent Attention Control (VCAC) module that leverages cross-view semantic and geometric reference from the source branch to yield aligned views from the target branch during the editing of 2D views. To validate the effectiveness of our method, we analyze 2D examples to demonstrate the improved consistency with the VCAC module. Further extensive quantitative and qualitative results in text-guided 3D scene editing indicate that our method achieves superior editing quality compared to state-of-the-art methods. We will make the complete codebase publicly available following the conclusion of the double-blind review process.
Authors: Wenxuan Guo, Yingping Liang, Zhiyu Pan, Ziheng Xi, Jianjiang Feng, Jie Zhou
Abstract: Gait recognition is a crucial biometric identification technique. Camera-based gait recognition has been widely applied in both research and industrial fields. LiDAR-based gait recognition has also begun to evolve most recently, due to the provision of 3D structural information. However, in certain applications, cameras fail to recognize persons, such as in low-light environments and long-distance recognition scenarios, where LiDARs work well. On the other hand, the deployment cost and complexity of LiDAR systems limit its wider application. Therefore, it is essential to consider cross-modality gait recognition between cameras and LiDARs for a broader range of applications. In this work, we propose the first cross-modality gait recognition framework between Camera and LiDAR, namely CL-Gait. It employs a two-stream network for feature embedding of both modalities. This poses a challenging recognition task due to the inherent matching between 3D and 2D data, exhibiting significant modality discrepancy. To align the feature spaces of the two modalities, i.e., camera silhouettes and LiDAR points, we propose a contrastive pre-training strategy to mitigate modality discrepancy. To make up for the absence of paired camera-LiDAR data for pre-training, we also introduce a strategy for generating data on a large scale. This strategy utilizes monocular depth estimated from single RGB images and virtual cameras to generate pseudo point clouds for contrastive pre-training. Extensive experiments show that the cross-modality gait recognition is very challenging but still contains potential and feasibility with our proposed model and pre-training strategy. To the best of our knowledge, this is the first work to address cross-modality gait recognition.
Authors: Zhiyuan Ma, Yuxiang Wei, Yabin Zhang, Xiangyu Zhu, Zhen Lei, Lei Zhang
Abstract: By leveraging the text-to-image diffusion priors, score distillation can synthesize 3D contents without paired text-3D training data. Instead of spending hours of online optimization per text prompt, recent studies have been focused on learning a text-to-3D generative network for amortizing multiple text-3D relations, which can synthesize 3D contents in seconds. However, existing score distillation methods are hard to scale up to a large amount of text prompts due to the difficulties in aligning pretrained diffusion prior with the distribution of rendered images from various text prompts. Current state-of-the-arts such as Variational Score Distillation finetune the pretrained diffusion model to minimize the noise prediction error so as to align the distributions, which are however unstable to train and will impair the model's comprehension capability to numerous text prompts. Based on the observation that the diffusion models tend to have lower noise prediction errors at earlier timesteps, we propose Asynchronous Score Distillation (ASD), which minimizes the noise prediction error by shifting the diffusion timestep to earlier ones. ASD is stable to train and can scale up to 100k prompts. It reduces the noise prediction error without changing the weights of pre-trained diffusion model, thus keeping its strong comprehension capability to prompts. We conduct extensive experiments across different 2D diffusion models, including Stable Diffusion and MVDream, and text-to-3D generators, including Hyper-iNGP, 3DConv-Net and Triplane-Transformer. The results demonstrate ASD's effectiveness in stable 3D generator training, high-quality 3D content synthesis, and its superior prompt-consistency, especially under large prompt corpus.
Authors: Hong Mo, Xiong Zhang, Jianchao Tan, Cheng Yang, Qiong Gu, Bo Hang, Wenqi Ren
Abstract: Multi-view counting (MVC) methods have shown their superiority over single-view counterparts, particularly in situations characterized by heavy occlusion and severe perspective distortions. However, hand-crafted heuristic features and identical camera layout requirements in conventional MVC methods limit their applicability and scalability in real-world scenarios.In this work, we propose a concise 3D MVC framework called \textbf{CountFormer}to elevate multi-view image-level features to a scene-level volume representation and estimate the 3D density map based on the volume features. By incorporating a camera encoding strategy, CountFormer successfully embeds camera parameters into the volume query and image-level features, enabling it to handle various camera layouts with significant differences.Furthermore, we introduce a feature lifting module capitalized on the attention mechanism to transform image-level features into a 3D volume representation for each camera view. Subsequently, the multi-view volume aggregation module attentively aggregates various multi-view volumes to create a comprehensive scene-level volume representation, allowing CountFormer to handle images captured by arbitrary dynamic camera layouts. The proposed method performs favorably against the state-of-the-art approaches across various widely used datasets, demonstrating its greater suitability for real-world deployment compared to conventional MVC frameworks.
Authors: Kaixin Xu, Zhe Wang, Chunyun Chen, Xue Geng, Jie Lin, Xulei Yang, Min Wu, Xiaoli Li, Weisi Lin
Abstract: Vision transformers have emerged as a promising alternative to convolutional neural networks for various image analysis tasks, offering comparable or superior performance. However, one significant drawback of ViTs is their resource-intensive nature, leading to increased memory footprint, computation complexity, and power consumption. To democratize this high-performance technology and make it more environmentally friendly, it is essential to compress ViT models, reducing their resource requirements while maintaining high performance. In this paper, we introduce a new block-structured pruning to address the resource-intensive issue for ViTs, offering a balanced trade-off between accuracy and hardware acceleration. Unlike unstructured pruning or channel-wise structured pruning, block pruning leverages the block-wise structure of linear layers, resulting in more efficient matrix multiplications. To optimize this pruning scheme, our paper proposes a novel hardware-aware learning objective that simultaneously maximizes speedup and minimizes power consumption during inference, tailored to the block sparsity structure. This objective eliminates the need for empirical look-up tables and focuses solely on reducing parametrized layer connections. Moreover, our paper provides a lightweight algorithm to achieve post-training pruning for ViTs, utilizing second-order Taylor approximation and empirical optimization to solve the proposed hardware-aware objective. Extensive experiments on ImageNet are conducted across various ViT architectures, including DeiT-B and DeiT-S, demonstrating competitive performance with other pruning methods and achieving a remarkable balance between accuracy preservation and power savings. Especially, we achieve up to 3.93x and 1.79x speedups on dedicated hardware and GPUs respectively for DeiT-B, and also observe an inference power reduction by 1.4x on real-world GPUs.
Authors: Pasquale De Marinis, Nicola Fanelli, Raffaele Scaringi, Emanuele Colonna, Giuseppe Fiameni, Gennaro Vessio, Giovanna Castellano
Abstract: We present Label Anything, an innovative neural network architecture designed for few-shot semantic segmentation (FSS) that demonstrates remarkable generalizability across multiple classes with minimal examples required per class. Diverging from traditional FSS methods that predominantly rely on masks for annotating support images, Label Anything introduces varied visual prompts -- points, bounding boxes, and masks -- thereby enhancing the framework's versatility and adaptability. Unique to our approach, Label Anything is engineered for end-to-end training across multi-class FSS scenarios, efficiently learning from diverse support set configurations without retraining. This approach enables a "universal" application to various FSS challenges, ranging from $1$-way $1$-shot to complex $N$-way $K$-shot configurations while remaining agnostic to the specific number of class examples. This innovative training strategy reduces computational requirements and substantially improves the model's adaptability and generalization across diverse segmentation tasks. Our comprehensive experimental validation, particularly achieving state-of-the-art results on the COCO-$20^i$ benchmark, underscores Label Anything's robust generalization and flexibility. The source code is publicly available at: https://github.com/pasqualedem/LabelAnything.
Authors: Bohan Li, Jiajun Deng, Wenyao Zhang, Zhujin Liang, Dalong Du, Xin Jin, Wenjun Zeng
Abstract: Camera-based 3D semantic scene completion (SSC) is pivotal for predicting complicated 3D layouts with limited 2D image observations. The existing mainstream solutions generally leverage temporal information by roughly stacking history frames to supplement the current frame, such straightforward temporal modeling inevitably diminishes valid clues and increases learning difficulty. To address this problem, we present HTCL, a novel Hierarchical Temporal Context Learning paradigm for improving camera-based semantic scene completion. The primary innovation of this work involves decomposing temporal context learning into two hierarchical steps: (a) cross-frame affinity measurement and (b) affinity-based dynamic refinement. Firstly, to separate critical relevant context from redundant information, we introduce the pattern affinity with scale-aware isolation and multiple independent learners for fine-grained contextual correspondence modeling. Subsequently, to dynamically compensate for incomplete observations, we adaptively refine the feature sampling locations based on initially identified locations with high affinity and their neighboring relevant regions. Our method ranks $1^{st}$ on the SemanticKITTI benchmark and even surpasses LiDAR-based methods in terms of mIoU on the OpenOccupancy benchmark. Our code is available on https://github.com/Arlo0o/HTCL.
Authors: Kaixin Xu, Qingtian Feng, Hao Chen, Zhe Wang, Xue Geng, Xulei Yang, Min Wu, Xiaoli Li, Weisi Lin
Abstract: Applying deep neural networks to 3D point cloud processing has attracted increasing attention due to its advanced performance in many areas, such as AR/VR, autonomous driving, and robotics. However, as neural network models and 3D point clouds expand in size, it becomes a crucial challenge to reduce the computational and memory overhead to meet latency and energy constraints in real-world applications. Although existing approaches have proposed to reduce both computational cost and memory footprint, most of them only address the spatial redundancy in inputs, i.e. removing the redundancy of background points in 3D data. In this paper, we propose a novel post-training weight pruning scheme for 3D object detection that is (1) orthogonal to all existing point cloud sparsifying methods, which determines redundant parameters in the pretrained model that lead to minimal distortion in both locality and confidence (detection distortion); and (2) a universal plug-and-play pruning framework that works with arbitrary 3D detection model. This framework aims to minimize detection distortion of network output to maximally maintain detection precision, by identifying layer-wise sparsity based on second-order Taylor approximation of the distortion. Albeit utilizing second-order information, we introduced a lightweight scheme to efficiently acquire Hessian information, and subsequently perform dynamic programming to solve the layer-wise sparsity. Extensive experiments on KITTI, Nuscenes and ONCE datasets demonstrate that our approach is able to maintain and even boost the detection precision on pruned model under noticeable computation reduction (FLOPs). Noticeably, we achieve over 3.89x, 3.72x FLOPs reduction on CenterPoint and PVRCNN model, respectively, without mAP decrease, significantly improving the state-of-the-art.
Authors: Nicola Messina, Jan Sedmidubsky, Fabrizio Falchi, Tom\'a\v{s} Rebok
Abstract: Pose-estimation methods enable extracting human motion from common videos in the structured form of 3D skeleton sequences. Despite great application opportunities, effective content-based access to such spatio-temporal motion data is a challenging problem. In this paper, we focus on the recently introduced text-motion retrieval tasks, which aim to search for database motions that are the most relevant to a specified natural-language textual description (text-to-motion) and vice-versa (motion-to-text). Despite recent efforts to explore these promising avenues, a primary challenge remains the insufficient data available to train robust text-motion models effectively. To address this issue, we propose to investigate joint-dataset learning - where we train on multiple text-motion datasets simultaneously - together with the introduction of a Cross-Consistent Contrastive Loss function (CCCL), which regularizes the learned text-motion common space by imposing uni-modal constraints that augment the representation ability of the trained network. To learn a proper motion representation, we also introduce a transformer-based motion encoder, called MoT++, which employs spatio-temporal attention to process sequences of skeleton data. We demonstrate the benefits of the proposed approaches on the widely-used KIT Motion-Language and HumanML3D datasets. We perform detailed experimentation on joint-dataset learning and cross-dataset scenarios, showing the effectiveness of each introduced module in a carefully conducted ablation study and, in turn, pointing out the limitations of state-of-the-art methods.
Authors: You Huang, Wenbin Lai, Jiayi Ji, Liujuan Cao, Shengchuan Zhang, Rongrong Ji
Abstract: The Segment Anything Model (SAM) has significantly advanced interactive segmentation but struggles with high-resolution images crucial for high-precision segmentation. This is primarily due to the quadratic space complexity of SAM-implemented attention and the length extrapolation issue in common global attention. This study proposes HRSAM that integrates Flash Attention and incorporates Plain, Shifted and newly proposed Cycle-scan Window (PSCWin) attention to address these issues. The shifted window attention is redesigned with padding to maintain consistent window sizes, enabling effective length extrapolation. The cycle-scan window attention adopts the recently developed State Space Models (SSMs) to ensure global information exchange with minimal computational overhead. Such window-based attention allows HRSAM to perform effective attention computations on scaled input images while maintaining low latency. Moreover, we further propose HRSAM++ that additionally employs a multi-scale strategy to enhance HRSAM's performance. The experiments on the high-precision segmentation datasets HQSeg44K and DAVIS show that high-resolution inputs enable the SAM-distilled HRSAM models to outperform the teacher model while maintaining lower latency. Compared to the SOTAs, HRSAM achieves a 1.56 improvement in interactive segmentation's NoC95 metric with only 31% of the latency. HRSAM++ further enhances the performance, achieving a 1.63 improvement in NoC95 with just 38% of the latency.
Authors: Shulei Qiu, Wanqi Yang, Ming Yang
Abstract: Our research focuses on few-shot fine-grained image classification, which faces two major challenges: appearance similarity of fine-grained objects and limited number of samples. To preserve the appearance details of images, traditional feature reconstruction networks usually enhance the representation ability of key features by spatial feature reconstruction and minimizing the reconstruction error. However, we find that relying solely on a single type of feature is insufficient for accurately capturing inter-class differences of fine-grained objects in scenarios with limited samples. In contrast, the introduction of channel features provides additional information dimensions, aiding in better understanding and distinguishing the inter-class differences of fine-grained objects. Therefore, in this paper, we design a new Hybrid Feature Collaborative Reconstruction Network (HFCR-Net) for few-shot fine-grained image classification, which includes a Hybrid Feature Fusion Process (HFFP) and a Hybrid Feature Reconstruction Process (HFRP). In HFRP, we fuse the channel features and the spatial features. Through dynamic weight adjustment, we aggregate the spatial dependencies between arbitrary two positions and the correlations between different channels of each image to increase the inter-class differences. Additionally, we introduce the reconstruction of channel dimension in HFRP. Through the collaborative reconstruction of channel dimension and spatial dimension, the inter-class differences are further increased in the process of support-to-query reconstruction, while the intra-class differences are reduced in the process of query-to-support reconstruction. Ultimately, our extensive experiments on three widely used fine-grained datasets demonstrate the effectiveness and superiority of our approach.
Authors: Ketan Kotwal, Ibrahim Ulucan, Gokhan Ozbulak, Janani Selliah, Sebastien Marcel
Abstract: With advancements in hardware, high-quality HMD devices are being developed by numerous companies, driving increased consumer interest in AR, VR, and MR applications. In this work, we present a new dataset, called VRBiom, of periocular videos acquired using a Virtual Reality headset. The VRBiom, targeted at biometric applications, consists of 900 short videos acquired from 25 individuals recorded in the NIR spectrum. These 10s long videos have been captured using the internal tracking cameras of Meta Quest Pro at 72 FPS. To encompass real-world variations, the dataset includes recordings under three gaze conditions: steady, moving, and partially closed eyes. We have also ensured an equal split of recordings without and with glasses to facilitate the analysis of eye-wear. These videos, characterized by non-frontal views of the eye and relatively low spatial resolutions (400 x 400), can be instrumental in advancing state-of-the-art research across various biometric applications. The VRBiom dataset can be utilized to evaluate, train, or adapt models for biometric use-cases such as iris and/or periocular recognition and associated sub-tasks such as detection and semantic segmentation. In addition to data from real individuals, we have included around 1100 PA constructed from 92 PA instruments. These PAIs fall into six categories constructed through combinations of print attacks (real and synthetic identities), fake 3D eyeballs, plastic eyes, and various types of masks and mannequins. These PA videos, combined with genuine (bona-fide) data, can be utilized to address concerns related to spoofing, which is a significant threat if these devices are to be used for authentication. The VRBiom dataset is publicly available for research purposes related to biometric applications only.
Authors: Haodong Chen, Haojian Huang, Junhao Dong, Mingzhe Zheng, Dian Shao
Abstract: Dynamic Facial Expression Recognition (DFER) is crucial for understanding human behavior. However, current methods exhibit limited performance mainly due to the scarcity of high-quality data, the insufficient utilization of facial dynamics, and the ambiguity of expression semantics, etc. To this end, we propose a novel framework, named Multi-modal Fine-grained CLIP for Dynamic Facial Expression Recognition with AdaptERs (FineCLIPER), incorporating the following novel designs: 1) To better distinguish between similar facial expressions, we extend the class labels to textual descriptions from both positive and negative aspects, and obtain supervision by calculating the cross-modal similarity based on the CLIP model; 2) Our FineCLIPER adopts a hierarchical manner to effectively mine useful cues from DFE videos. Specifically, besides directly embedding video frames as input (low semantic level), we propose to extract the face segmentation masks and landmarks based on each frame (middle semantic level) and utilize the Multi-modal Large Language Model (MLLM) to further generate detailed descriptions of facial changes across frames with designed prompts (high semantic level). Additionally, we also adopt Parameter-Efficient Fine-Tuning (PEFT) to enable efficient adaptation of large pre-trained models (i.e., CLIP) for this task. Our FineCLIPER achieves SOTA performance on the DFEW, FERV39k, and MAFW datasets in both supervised and zero-shot settings with few tunable parameters. Analysis and ablation studies further validate its effectiveness.
Authors: Jingjing Ren, Wenbo Li, Haoyu Chen, Renjing Pei, Bin Shao, Yong Guo, Long Peng, Fenglong Song, Lei Zhu
Abstract: Ultra-high-resolution image generation poses great challenges, such as increased semantic planning complexity and detail synthesis difficulties, alongside substantial training resource demands. We present UltraPixel, a novel architecture utilizing cascade diffusion models to generate high-quality images at multiple resolutions (\textit{e.g.}, 1K to 6K) within a single model, while maintaining computational efficiency. UltraPixel leverages semantics-rich representations of lower-resolution images in the later denoising stage to guide the whole generation of highly detailed high-resolution images, significantly reducing complexity. Furthermore, we introduce implicit neural representations for continuous upsampling and scale-aware normalization layers adaptable to various resolutions. Notably, both low- and high-resolution processes are performed in the most compact space, sharing the majority of parameters with less than 3$\%$ additional parameters for high-resolution outputs, largely enhancing training and inference efficiency. Our model achieves fast training with reduced data requirements, producing photo-realistic high-resolution images and demonstrating state-of-the-art performance in extensive experiments.
Authors: Jintu Zheng, YI Ding, Qizhe Liu, Yi Cao, Ying Hu, Zenan Wang
Abstract: Traditional fluorescence staining is phototoxic to live cells, slow, and expensive; thus, the subcellular structure prediction (SSP) from transmitted light (TL) images is emerging as a label-free, faster, low-cost alternative. However, existing approaches utilize 3D networks for one-to-one voxel level dense prediction, which necessitates a frequent and time-consuming Z-axis imaging process. Moreover, 3D convolutions inevitably lead to significant computation and GPU memory overhead. Therefore, we propose an efficient framework, SparseSSP, predicting fluorescent intensities within the target voxel grid in an efficient paradigm instead of relying entirely on 3D topologies. In particular, SparseSSP makes two pivotal improvements to prior works. First, SparseSSP introduces a one-to-many voxel mapping paradigm, which permits the sparse TL slices to reconstruct the subcellular structure. Secondly, we propose a hybrid dimensions topology, which folds the Z-axis information into channel features, enabling the 2D network layers to tackle SSP under low computational cost. We conduct extensive experiments to validate the effectiveness and advantages of SparseSSP on diverse sparse imaging ratios, and our approach achieves a leading performance compared to pure 3D topologies. SparseSSP reduces imaging frequencies compared to previous dense-view SSP (i.e., the number of imaging is reduced up to 87.5% at most), which is significant in visualizing rapid biological dynamics on low-cost devices and samples.
Authors: Zihao Huang, ShouKang Hu, Guangcong Wang, Tianqi Liu, Yuhang Zang, Zhiguo Cao, Wei Li, Ziwei Liu
Abstract: Existing human datasets for avatar creation are typically limited to laboratory environments, wherein high-quality annotations (e.g., SMPL estimation from 3D scans or multi-view images) can be ideally provided. However, their annotating requirements are impractical for real-world images or videos, posing challenges toward real-world applications on current avatar creation methods. To this end, we propose the WildAvatar dataset, a web-scale in-the-wild human avatar creation dataset extracted from YouTube, with $10,000+$ different human subjects and scenes. WildAvatar is at least $10\times$ richer than previous datasets for 3D human avatar creation. We evaluate several state-of-the-art avatar creation methods on our dataset, highlighting the unexplored challenges in real-world applications on avatar creation. We also demonstrate the potential for generalizability of avatar creation methods, when provided with data at scale. We will publicly release our data source links and annotations, to push forward 3D human avatar creation and other related fields for real-world applications.
Authors: Wenpu Li, Pian Wan, Peng Wang, Jinhang Li, Yi Zhou, Peidong Liu
Abstract: Neural implicit representation of visual scenes has attracted a lot of attention in recent research of computer vision and graphics. Most prior methods focus on how to reconstruct 3D scene representation from a set of images. In this work, we demonstrate the possibility to recover the neural radiance fields (NeRF) from a single blurry image and its corresponding event stream. We model the camera motion with a cubic B-Spline in SE(3) space. Both the blurry image and the brightness change within a time interval, can then be synthesized from the 3D scene representation given the 6-DoF poses interpolated from the cubic B-Spline. Our method can jointly learn both the implicit neural scene representation and recover the camera motion by minimizing the differences between the synthesized data and the real measurements without pre-computed camera poses from COLMAP. We evaluate the proposed method with both synthetic and real datasets. The experimental results demonstrate that we are able to render view-consistent latent sharp images from the learned NeRF and bring a blurry image alive in high quality. Code and data are available at https://github.com/WU-CVGL/BeNeRF.
Authors: Yihong Cao, Jiaming Zhang, Hao Shi, Kunyu Peng, Yuhongxuan Zhang, Hui Zhang, Rainer Stiefelhagen, Kailun Yang
Abstract: Panoramic images can broaden the Field of View (FoV), occlusion-aware prediction can deepen the understanding of the scene, and domain adaptation can transfer across viewing domains. In this work, we introduce a novel task, Occlusion-Aware Seamless Segmentation (OASS), which simultaneously tackles all these three challenges. For benchmarking OASS, we establish a new human-annotated dataset for Blending Panoramic Amodal Seamless Segmentation, i.e., BlendPASS. Besides, we propose the first solution UnmaskFormer, aiming at unmasking the narrow FoV, occlusions, and domain gaps all at once. Specifically, UnmaskFormer includes the crucial designs of Unmasking Attention (UA) and Amodal-oriented Mix (AoMix). Our method achieves state-of-the-art performance on the BlendPASS dataset, reaching a remarkable mAPQ of 26.58% and mIoU of 43.66%. On public panoramic semantic segmentation datasets, i.e., SynPASS and DensePASS, our method outperforms previous methods and obtains 45.34% and 48.08% in mIoU, respectively. The fresh BlendPASS dataset and our source code will be made publicly available at https://github.com/yihong-97/OASS.
Authors: Katja L\"owenstein, Johanna Rehrl, Anja Schuster, Michael Gadermayr
Abstract: The in vitro scratch assay is a widely used assay in cell biology to assess the rate of wound closure related to a variety of therapeutic interventions. While manual measurement is subjective and vulnerable to intra- and interobserver variability, computer-based tools are theoretically objective, but in practice often contain parameters which are manually adjusted (individually per image or data set) and thereby provide a source for subjectivity. Modern deep learning approaches typically require large annotated training data which complicates instant applicability. In this paper, we make use of the segment anything model, a deep foundation model based on interactive point-prompts, which enables class-agnostic segmentation without tuning the network's parameters based on domain specific training data. The proposed method clearly outperformed a semi-objective baseline method that required manual inspection and, if necessary, adjustment of parameters per image. Even though the point prompts of the proposed approach are theoretically also a source for subjectivity, results attested very low intra- and interobserver variability, even compared to manual segmentation of domain experts.
Authors: Adnen Abdessaied, Lei Shi, Andreas Bulling
Abstract: We present MST-MIXER - a novel video dialog model operating over a generic multi-modal state tracking scheme. Current models that claim to perform multi-modal state tracking fall short of two major aspects: (1) They either track only one modality (mostly the visual input) or (2) they target synthetic datasets that do not reflect the complexity of real-world in the wild scenarios. Our model addresses these two limitations in an attempt to close this crucial research gap. Specifically, MST-MIXER first tracks the most important constituents of each input modality. Then, it predicts the missing underlying structure of the selected constituents of each modality by learning local latent graphs using a novel multi-modal graph structure learning method. Subsequently, the learned local graphs and features are parsed together to form a global graph operating on the mix of all modalities which further refines its structure and node embeddings. Finally, the fine-grained graph node features are used to enhance the hidden states of the backbone Vision-Language Model (VLM). MST-MIXER achieves new state-of-the-art results on five challenging benchmarks.
Authors: Ali Akin, Habil Kalkan
Abstract: Traffic accidents, causing millions of deaths and billions of dollars in economic losses each year globally, have become a significant issue. One of the main causes of these accidents is drivers being sleepy or fatigued. Recently, various studies have focused on detecting drivers' sleep/wake states using camera-based solutions that do not require physical contact with the driver, thereby enhancing ease of use. In this study, besides the eye blink frequency, a driver adaptive eye blink behavior feature set have been evaluated to detect the fatigue status. It is observed from the results that behavior of eye blink carries useful information on fatigue detection. The developed image-based system provides a solution that can work adaptively to the physical characteristics of the drivers and their positions in the vehicle
Authors: Baijiong Lin, Weisen Jiang, Pengguang Chen, Yu Zhang, Shu Liu, Ying-Cong Chen
Abstract: Multi-task dense scene understanding, which learns a model for multiple dense prediction tasks, has a wide range of application scenarios. Modeling long-range dependency and enhancing cross-task interactions are crucial to multi-task dense prediction. In this paper, we propose MTMamba, a novel Mamba-based architecture for multi-task scene understanding. It contains two types of core blocks: self-task Mamba (STM) block and cross-task Mamba (CTM) block. STM handles long-range dependency by leveraging Mamba, while CTM explicitly models task interactions to facilitate information exchange across tasks. Experiments on NYUDv2 and PASCAL-Context datasets demonstrate the superior performance of MTMamba over Transformer-based and CNN-based methods. Notably, on the PASCAL-Context dataset, MTMamba achieves improvements of +2.08, +5.01, and +4.90 over the previous best method in the tasks of semantic segmentation, human parsing, and object boundary detection, respectively. The code is available at \url{https://github.com/EnVision-Research/MTMamba}.
Authors: Jiarui Xing, Nivetha Jayakumar, Nian Wu, Yu Wang, Frederick H. Epstein, Miaomiao Zhang
Abstract: Motion and deformation analysis of cardiac magnetic resonance (CMR) imaging videos is crucial for assessing myocardial strain of patients with abnormal heart functions. Recent advances in deep learning-based image registration algorithms have shown promising results in predicting motion fields from routinely acquired CMR sequences. However, their accuracy often diminishes in regions with subtle appearance change, with errors propagating over time. Advanced imaging techniques, such as displacement encoding with stimulated echoes (DENSE) CMR, offer highly accurate and reproducible motion data but require additional image acquisition, which poses challenges in busy clinical flows. In this paper, we introduce a novel Latent Motion Diffusion model (LaMoD) to predict highly accurate DENSE motions from standard CMR videos. More specifically, our method first employs an encoder from a pre-trained registration network that learns latent motion features (also considered as deformation-based shape features) from image sequences. Supervised by the ground-truth motion provided by DENSE, LaMoD then leverages a probabilistic latent diffusion model to reconstruct accurate motion from these extracted features. Experimental results demonstrate that our proposed method, LaMoD, significantly improves the accuracy of motion analysis in standard CMR images; hence improving myocardial strain analysis in clinical settings for cardiac patients. Our code will be publicly available on upon acceptance.
Authors: Zhiyu Long, Xingyou Liu, Jiaqi Qiao, Zhi Li
Abstract: Sign language is a visual language used by the deaf and dumb community to communicate. However, for most recognition methods based on monocular cameras, the recognition accuracy is low and the robustness is poor. Even if the effect is good on some data, it may perform poorly in other data with different interference due to the inability to extract effective features. To solve these problems, we propose a sign language recognition network that integrates skeleton features of hands and facial expression. Among this, we propose a hand skeleton feature extraction based on coordinate transformation to describe the shape of the hand more accurately. Moreover, by incorporating facial expression information, the accuracy and robustness of sign language recognition are finally improved, which was verified on A Dataset for Argentinian Sign Language and SEU's Chinese Sign Language Recognition Database (SEUCSLRD).
Authors: Ayane Tajima, Satoshi Ono
Abstract: Research has shown that deep neural networks (DNNs) have vulnerabilities that can lead to the misrecognition of Adversarial Examples (AEs) with specifically designed perturbations. Various adversarial attack methods have been proposed to detect vulnerabilities under hard-label black box (HL-BB) conditions in the absence of loss gradients and confidence scores.However, these methods fall into local solutions because they search only local regions of the search space. Therefore, this study proposes an adversarial attack method named EvolBA to generate AEs using Covariance Matrix Adaptation Evolution Strategy (CMA-ES) under the HL-BB condition, where only a class label predicted by the target DNN model is available. Inspired by formula-driven supervised learning, the proposed method introduces domain-independent operators for the initialization process and a jump that enhances search exploration. Experimental results confirmed that the proposed method could determine AEs with smaller perturbations than previous methods in images where the previous methods have difficulty.
Authors: Jian Ma, Yonglin Deng, Chen Chen, Haonan Lu, Zhenyu Yang
Abstract: Posters play a crucial role in marketing and advertising, contributing significantly to industrial design by enhancing visual communication and brand visibility. With recent advances in controllable text-to-image diffusion models, more concise research is now focusing on rendering text within synthetic images. Despite improvements in text rendering accuracy, the field of end-to-end poster generation remains underexplored. This complex task involves striking a balance between text rendering accuracy and automated layout to produce high-resolution images with variable aspect ratios. To tackle this challenge, we propose an end-to-end text rendering framework employing a triple cross-attention mechanism rooted in align learning, designed to create precise poster text within detailed contextual backgrounds. Additionally, we introduce a high-resolution dataset that exceeds 1024 pixels in image resolution. Our approach leverages the SDXL architecture. Extensive experiments validate the ability of our method to generate poster images featuring intricate and contextually rich backgrounds. Codes will be available at https://github.com/OPPO-Mente-Lab/GlyphDraw2.
Authors: Sufen Ren, Yule Hu, Shengchao Chen, Guanjun Wang
Abstract: Medical image classification plays a crucial role in computer-aided clinical diagnosis. While deep learning techniques have significantly enhanced efficiency and reduced costs, the privacy-sensitive nature of medical imaging data complicates centralized storage and model training. Furthermore, low-resource healthcare organizations face challenges related to communication overhead and efficiency due to increasing data and model scales. This paper proposes a novel privacy-preserving medical image classification framework based on federated learning to address these issues, named FedMIC. The framework enables healthcare organizations to learn from both global and local knowledge, enhancing local representation of private data despite statistical heterogeneity. It provides customized models for organizations with diverse data distributions while minimizing communication overhead and improving efficiency without compromising performance. Our FedMIC enhances robustness and practical applicability under resource-constrained conditions. We demonstrate FedMIC's effectiveness using four public medical image datasets for classical medical image classification tasks.
Authors: Huiyu Gao, Jiahao Ma, David Ahmedt-Aristizabal, Chuong Nguyen, Miaomiao Liu
Abstract: This paper tackles the problem of novel view audio-visual synthesis along an arbitrary trajectory in an indoor scene, given the audio-video recordings from other known trajectories of the scene. Existing methods often overlook the effect of room geometry, particularly wall occlusion to sound propagation, making them less accurate in multi-room environments. In this work, we propose a new approach called Scene Occlusion-aware Acoustic Field (SOAF) for accurate sound generation. Our approach derives a prior for sound energy field using distance-aware parametric sound-propagation modelling and then transforms it based on scene transmittance learned from the input video. We extract features from the local acoustic field centred around the receiver using a Fibonacci Sphere to generate binaural audio for novel views with a direction-aware attention mechanism. Extensive experiments on the real dataset~\emph{RWAVS} and the synthetic dataset~\emph{SoundSpaces} demonstrate that our method outperforms previous state-of-the-art techniques in audio generation. Project page: https://github.com/huiyu-gao/SOAF/.
Authors: Haoru Wang, Wentao Zhu, Luyi Miao, Yishu Xu, Feng Gao, Qi Tian, Yizhou Wang
Abstract: Human motion generation is a critical task with a wide range of applications. Achieving high realism in generated motions requires naturalness, smoothness, and plausibility. Despite rapid advancements in the field, current generation methods often fall short of these goals. Furthermore, existing evaluation metrics typically rely on ground-truth-based errors, simple heuristics, or distribution distances, which do not align well with human perceptions of motion quality. In this work, we propose a data-driven approach to bridge this gap by introducing a large-scale human perceptual evaluation dataset, MotionPercept, and a human motion critic model, MotionCritic, that capture human perceptual preferences. Our critic model offers a more accurate metric for assessing motion quality and could be readily integrated into the motion generation pipeline to enhance generation quality. Extensive experiments demonstrate the effectiveness of our approach in both evaluating and improving the quality of generated human motions by aligning with human perceptions. Code and data are publicly available at https://motioncritic.github.io/.
Authors: Yangyang Xiang, Nannan Wu, Li Yu, Xin Yang, Kwang-Ting Cheng, Zengqiang Yan
Abstract: Federated learning has emerged as a compelling paradigm for medical image segmentation, particularly in light of increasing privacy concerns. However, most of the existing research relies on relatively stringent assumptions regarding the uniformity and completeness of annotations across clients. Contrary to this, this paper highlights a prevalent challenge in medical practice: incomplete annotations. Such annotations can introduce incorrectly labeled pixels, potentially undermining the performance of neural networks in supervised learning. To tackle this issue, we introduce a novel solution, named FedIA. Our insight is to conceptualize incomplete annotations as noisy data (\textit{i.e.}, low-quality data), with a focus on mitigating their adverse effects. We begin by evaluating the completeness of annotations at the client level using a designed indicator. Subsequently, we enhance the influence of clients with more comprehensive annotations and implement corrections for incomplete ones, thereby ensuring that models are trained on accurate data. Our method's effectiveness is validated through its superior performance on two extensively used medical image segmentation datasets, outperforming existing solutions. The code is available at https://github.com/HUSTxyy/FedIA.
Authors: Minghao Zhou, Hong Wang, Yefeng Zheng, Deyu Meng
Abstract: Feature upsampling is a fundamental and indispensable ingredient of almost all current network structures for image segmentation tasks. Recently, a popular similarity-based feature upsampling pipeline has been proposed, which utilizes a high-resolution feature as guidance to help upsample the low-resolution deep feature based on their local similarity. Albeit achieving promising performance, this pipeline has specific limitations: 1) HR query and LR key features are not well aligned; 2) the similarity between query-key features is computed based on the fixed inner product form; 3) neighbor selection is coarsely operated on LR features, resulting in mosaic artifacts. These shortcomings make the existing methods along this pipeline primarily applicable to hierarchical network architectures with iterative features as guidance and they are not readily extended to a broader range of structures, especially for a direct high-ratio upsampling. Against the issues, we meticulously optimize every methodological design. Specifically, we firstly propose an explicitly controllable query-key feature alignment from both semantic-aware and detail-aware perspectives, and then construct a parameterized paired central difference convolution block for flexibly calculating the similarity between the well-aligned query-key features. Besides, we develop a fine-grained neighbor selection strategy on HR features, which is simple yet effective for alleviating mosaic artifacts. Based on these careful designs, we systematically construct a refreshed similarity-based feature upsampling framework named ReSFU. Extensive experiments substantiate that our proposed ReSFU is finely applicable to various types of architectures in a direct high-ratio upsampling manner, and consistently achieves satisfactory performance on different segmentation applications, showing superior generality and ease of deployment.
Authors: Junsung Park, Kyungmin Kim, Hyunjung Shim
Abstract: Existing LiDAR semantic segmentation methods often struggle with performance declines in adverse weather conditions. Previous research has addressed this issue by simulating adverse weather or employing universal data augmentation during training. However, these methods lack a detailed analysis and understanding of how adverse weather negatively affects LiDAR semantic segmentation performance. Motivated by this issue, we identified key factors of adverse weather and conducted a toy experiment to pinpoint the main causes of performance degradation: (1) Geometric perturbation due to refraction caused by fog or droplets in the air and (2) Point drop due to energy absorption and occlusions. Based on these findings, we propose new strategic data augmentation techniques. First, we introduced a Selective Jittering (SJ) that jitters points in the random range of depth (or angle) to mimic geometric perturbation. Additionally, we developed a Learnable Point Drop (LPD) to learn vulnerable erase patterns with Deep Q-Learning Network to approximate the point drop phenomenon from adverse weather conditions. Without precise weather simulation, these techniques strengthen the LiDAR semantic segmentation model by exposing it to vulnerable conditions identified by our data-centric analysis. Experimental results confirmed the suitability of the proposed data augmentation methods for enhancing robustness against adverse weather conditions. Our method attains a remarkable 39.5 mIoU on the SemanticKITTI-to-SemanticSTF benchmark, surpassing the previous state-of-the-art by over 5.4%p, tripling the improvement over the baseline compared to previous methods achieved.
Authors: Anxhelo Diko, Danilo Avola, Bardh Prenkaj, Federico Fontana, Luigi Cinque
Abstract: Action anticipation is the task of forecasting future activity from a partially observed sequence of events. However, this task is exposed to intrinsic future uncertainty and the difficulty of reasoning upon interconnected actions. Unlike previous works that focus on extrapolating better visual and temporal information, we concentrate on learning action representations that are aware of their semantic interconnectivity based on prototypical action patterns and contextual co-occurrences. To this end, we propose the novel Semantically Guided Representation Learning (S-GEAR) framework. S-GEAR learns visual action prototypes and leverages language models to structure their relationship, inducing semanticity. To gather insights on S-GEAR's effectiveness, we test it on four action anticipation benchmarks, obtaining improved results compared to previous works: +3.5, +2.7, and +3.5 absolute points on Top-1 Accuracy on Epic-Kitchen 55, EGTEA Gaze+ and 50 Salads, respectively, and +0.8 on Top-5 Recall on Epic-Kitchens 100. We further observe that S-GEAR effectively transfers the geometric associations between actions from language to visual prototypes. Finally, S-GEAR opens new research frontiers in anticipation tasks by demonstrating the intricate impact of action semantic interconnectivity.
Authors: Guozhen Zhang, Chunxu Liu, Yutao Cui, Xiaotong Zhao, Kai Ma, Limin Wang
Abstract: Inter-frame modeling is pivotal in generating intermediate frames for video frame interpolation (VFI). Current approaches predominantly rely on convolution or attention-based models, which often either lack sufficient receptive fields or entail significant computational overheads. Recently, Selective State Space Models (S6) have emerged, tailored specifically for long sequence modeling, offering both linear complexity and data-dependent modeling capabilities. In this paper, we propose VFIMamba, a novel frame interpolation method for efficient and dynamic inter-frame modeling by harnessing the S6 model. Our approach introduces the Mixed-SSM Block (MSB), which initially rearranges tokens from adjacent frames in an interleaved fashion and subsequently applies multi-directional S6 modeling. This design facilitates the efficient transmission of information across frames while upholding linear complexity. Furthermore, we introduce a novel curriculum learning strategy that progressively cultivates proficiency in modeling inter-frame dynamics across varying motion magnitudes, fully unleashing the potential of the S6 model. Experimental findings showcase that our method attains state-of-the-art performance across diverse benchmarks, particularly excelling in high-resolution scenarios. In particular, on the X-TEST dataset, VFIMamba demonstrates a noteworthy improvement of 0.80 dB for 4K frames and 0.96 dB for 2K frames.
Authors: Dewei Zhou, You Li, Fan Ma, Zongxin Yang, Yi Yang
Abstract: We introduce the Multi-Instance Generation (MIG) task, which focuses on generating multiple instances within a single image, each accurately placed at predefined positions with attributes such as category, color, and shape, strictly following user specifications. MIG faces three main challenges: avoiding attribute leakage between instances, supporting diverse instance descriptions, and maintaining consistency in iterative generation. To address attribute leakage, we propose the Multi-Instance Generation Controller (MIGC). MIGC generates multiple instances through a divide-and-conquer strategy, breaking down multi-instance shading into single-instance tasks with singular attributes, later integrated. To provide more types of instance descriptions, we developed MIGC++. MIGC++ allows attribute control through text \& images and position control through boxes \& masks. Lastly, we introduced the Consistent-MIG algorithm to enhance the iterative MIG ability of MIGC and MIGC++. This algorithm ensures consistency in unmodified regions during the addition, deletion, or modification of instances, and preserves the identity of instances when their attributes are changed. We introduce the COCO-MIG and Multimodal-MIG benchmarks to evaluate these methods. Extensive experiments on these benchmarks, along with the COCO-Position benchmark and DrawBench, demonstrate that our methods substantially outperform existing techniques, maintaining precise control over aspects including position, attribute, and quantity. Project page: https://github.com/limuloo/MIGC.
Authors: Yi Zhang, Ke Yu, Siqi Wu, Zhihai He
Abstract: In this paper, we propose Conceptual Codebook Learning (CoCoLe), a novel fine-tuning method for vision-language models (VLMs) to address the challenge of improving the generalization capability of VLMs while fine-tuning them on downstream tasks in a few-shot setting. We recognize that visual concepts, such as textures, shapes, and colors are naturally transferable across domains and play a crucial role in generalization tasks. Motivated by this interesting finding, we learn a conceptual codebook consisting of visual concepts as keys and conceptual prompts as values, which serves as a link between the image encoder's outputs and the text encoder's inputs. Specifically, for a given image, we leverage the codebook to identify the most relevant conceptual prompts associated with the class embeddings to perform the classification. Additionally, we incorporate a handcrafted concept cache as a regularization to alleviate the overfitting issues in low-shot scenarios. We observe that this conceptual codebook learning method is able to achieve enhanced alignment between visual and linguistic modalities. Extensive experimental results demonstrate that our CoCoLe method remarkably outperforms the existing state-of-the-art methods across various evaluation settings, including base-to-new generalization, cross-dataset evaluation, and domain generalization tasks. Detailed ablation studies further confirm the efficacy of each component in CoCoLe.
Authors: Hozaifa Kassab, Mohamed Bahaa, Ali Hamdi
Abstract: Facial Expression Recognition (FER) is vital for understanding interpersonal communication. However, existing classification methods often face challenges such as vulnerability to noise, imbalanced datasets, overfitting, and generalization issues. In this paper, we propose GCF, a novel approach that utilizes Graph Convolutional Networks for FER. GCF integrates Convolutional Neural Networks (CNNs) for feature extraction, using either custom architectures or pretrained models. The extracted visual features are then represented on a graph, enhancing local CNN features with global features via a Graph Convolutional Neural Network layer. We evaluate GCF on benchmark datasets including CK+, JAFFE, and FERG. The results show that GCF significantly improves performance over state-of-the-art methods. For example, GCF enhances the accuracy of ResNet18 from 92% to 98% on CK+, from 66% to 89% on JAFFE, and from 94% to 100% on FERG. Similarly, GCF improves the accuracy of VGG16 from 89% to 97% on CK+, from 72% to 92% on JAFFE, and from 96% to 99.49% on FERG. We provide a comprehensive analysis of our approach, demonstrating its effectiveness in capturing nuanced facial expressions. By integrating graph convolutions with CNNs, GCF significantly advances FER, offering improved accuracy and robustness in real-world applications.
Authors: Antoine Deckyvere, Anthony Cioppa, Silvio Giancola, Bernard Ghanem, Marc Van Droogenbroeck
Abstract: Slow-motion replays provide a thrilling perspective on pivotal moments within sports games, offering a fresh and captivating visual experience. However, capturing slow-motion footage typically demands high-tech, expensive cameras and infrastructures. Deep learning Video Frame Interpolation (VFI) techniques have emerged as a promising avenue, capable of generating high-speed footage from regular camera feeds. Moreover, the utilization of event-based cameras has recently gathered attention as they provide valuable motion information between frames, further enhancing the VFI performances. In this work, we present a first investigation of event-based VFI models for generating sports slow-motion videos. Particularly, we design and implement a bi-camera recording setup, including an RGB and an event-based camera to capture sports videos, to temporally align and spatially register both cameras. Our experimental validation demonstrates that TimeLens, an off-the-shelf event-based VFI model, can effectively generate slow-motion footage for sports videos. This first investigation underscores the practical utility of event-based cameras in producing sports slow-motion content and lays the groundwork for future research endeavors in this domain.
Authors: Kepan Nan, Rui Xie, Penghao Zhou, Tiehan Fan, Zhenheng Yang, Zhijie Chen, Xiang Li, Jian Yang, Ying Tai
Abstract: Text-to-video (T2V) generation has recently garnered significant attention thanks to the large multi-modality model Sora. However, T2V generation still faces two important challenges: 1) Lacking a precise open sourced high-quality dataset. The previous popular video datasets, e.g. WebVid-10M and Panda-70M, are either with low quality or too large for most research institutions. Therefore, it is challenging but crucial to collect a precise high-quality text-video pairs for T2V generation. 2) Ignoring to fully utilize textual information. Recent T2V methods have focused on vision transformers, using a simple cross attention module for video generation, which falls short of thoroughly extracting semantic information from text prompt. To address these issues, we introduce OpenVid-1M, a precise high-quality dataset with expressive captions. This open-scenario dataset contains over 1 million text-video pairs, facilitating research on T2V generation. Furthermore, we curate 433K 1080p videos from OpenVid-1M to create OpenVidHD-0.4M, advancing high-definition video generation. Additionally, we propose a novel Multi-modal Video Diffusion Transformer (MVDiT) capable of mining both structure information from visual tokens and semantic information from text tokens. Extensive experiments and ablation studies verify the superiority of OpenVid-1M over previous datasets and the effectiveness of our MVDiT.
Authors: Zhiqi Zhao, Chang Wu, Xiaotong Kong, Zejie Lv, Xiaoqi Du, Qiyan Li
Abstract: Simultaneous Localization and Mapping (SLAM) has become a critical technology for intelligent transportation systems and autonomous robots and is widely used in autonomous driving. However, traditional manual feature-based methods in challenging lighting environments make it difficult to ensure robustness and accuracy. Some deep learning-based methods show potential but still have significant drawbacks. To address this problem, we propose a novel hybrid system for visual SLAM based on the LightGlue deep learning network. It uses deep local feature descriptors to replace traditional hand-crafted features and a more efficient and accurate deep network to achieve fast and precise feature matching. Thus, we use the robustness of deep learning to improve the whole system. We have combined traditional geometry-based approaches to introduce a complete visual SLAM system for monocular, binocular, and RGB-D sensors. We thoroughly tested the proposed system on four public datasets: KITTI, EuRoC, TUM, and 4Season, as well as on actual campus scenes. The experimental results show that the proposed method exhibits better accuracy and robustness in adapting to low-light and strongly light-varying environments than traditional manual features and deep learning-based methods. It can also run on GPU in real time.
Authors: Xu Yin, Fei Pan, Guoyuan An, Yuchi Huo, Zixuan Xie, Sung-Eui Yoon
Abstract: Existing open-set recognition (OSR) studies typically assume that each image contains only one class label, and the unknown test set (negative) has a disjoint label space from the known test set (positive), a scenario termed full-label shift. This paper introduces the mixed OSR problem, where test images contain multiple class semantics, with known and unknown classes co-occurring in negatives, leading to a more challenging super-label shift. Addressing the mixed OSR requires classification models to accurately distinguish different class semantics within images and measure their "knowness". In this study, we propose the OpenSlot framework, built upon object-centric learning. OpenSlot utilizes slot features to represent diverse class semantics and produce class predictions. Through our proposed anti-noise-slot (ANS) technique, we mitigate the impact of noise (invalid and background) slots during classification training, effectively addressing the semantic misalignment between class predictions and the ground truth. We conduct extensive experiments with OpenSlot on mixed & conventional OSR benchmarks. Without elaborate designs, OpenSlot not only exceeds existing OSR studies in detecting super-label shifts across single & multi-label mixed OSR tasks but also achieves state-of-the-art performance on conventional benchmarks. Remarkably, our method can localize class objects without using bounding boxes during training. The competitive performance in open-set object detection demonstrates OpenSlot's ability to explicitly explain label shifts and benefits in computational efficiency and generalization.
Authors: Shuangliang Li
Abstract: Nowadays, most of the hyperspectral image (HSI) fusion experiments are based on simulated datasets to compare different fusion methods. However, most of the spectral response functions and spatial downsampling functions used to create the simulated datasets are not entirely accurate, resulting in deviations in spatial and spectral features between the generated images for fusion and the real images for fusion. This reduces the credibility of the fusion algorithm, causing unfairness in the comparison between different algorithms and hindering the development of the field of hyperspectral image fusion. Therefore, we release a real HSI/MSI/PAN image dataset to promote the development of the field of hyperspectral image fusion. These three images are spatially registered, meaning fusion can be performed between HSI and MSI, HSI and PAN image, MSI and PAN image, as well as among HSI, MSI, and PAN image. This real dataset could be available at https://aistudio.baidu.com/datasetdetail/281612. The related code to process the data could be available at https://github.com/rs-lsl/CSSNet.
URLs: https://aistudio.baidu.com/datasetdetail/281612., https://github.com/rs-lsl/CSSNet.
Authors: Sayan Nag, Koustava Goswami, Srikrishna Karanam
Abstract: Referring Expression Segmentation (RES) aims to provide a segmentation mask of the target object in an image referred to by the text (i.e., referring expression). Existing methods require large-scale mask annotations. Moreover, such approaches do not generalize well to unseen/zero-shot scenarios. To address the aforementioned issues, we propose a weakly-supervised bootstrapping architecture for RES with several new algorithmic innovations. To the best of our knowledge, ours is the first approach that considers only a fraction of both mask and box annotations (shown in Figure 1 and Table 1) for training. To enable principled training of models in such low-annotation settings, improve image-text region-level alignment, and further enhance spatial localization of the target object in the image, we propose Cross-modal Fusion with Attention Consistency module. For automatic pseudo-labeling of unlabeled samples, we introduce a novel Mask Validity Filtering routine based on a spatially aware zero-shot proposal scoring approach. Extensive experiments show that with just 30% annotations, our model SafaRi achieves 59.31 and 48.26 mIoUs as compared to 58.93 and 48.19 mIoUs obtained by the fully-supervised SOTA method SeqTR respectively on RefCOCO+@testA and RefCOCO+testB datasets. SafaRi also outperforms SeqTR by 11.7% (on RefCOCO+testA) and 19.6% (on RefCOCO+testB) in a fully-supervised setting and demonstrates strong generalization capabilities in unseen/zero-shot tasks.
Authors: Wentong Li, Yuqian Yuan, Jian Liu, Dongqi Tang, Song Wang, Jianke Zhu, Lei Zhang
Abstract: The visual projector serves as an essential bridge between the visual encoder and the Large Language Model (LLM) in a Multimodal LLM (MLLM). Typically, MLLMs adopt a simple MLP to preserve all visual contexts via one-to-one transformation. However, the visual tokens are redundant and can be considerably increased when dealing with high-resolution images, impairing the efficiency of MLLMs significantly. Some recent works have introduced resampler or abstractor to reduce the number of resulting visual tokens. Unfortunately, they fail to capture finer details and undermine the visual reasoning capabilities of MLLMs. In this work, we propose a novel visual projector, which adopts a coarse-to-fine scheme to inject the enriched characteristics to generate the condensed visual tokens. In specific, we first interpolate the visual features as a low-resolution point query, providing the overall visual representation as the foundation. Then, we introduce a region-to-point injection module that utilizes high-resolution, multi-level region-based cues as fine-grained reference keys and values, allowing them to be fully absorbed within the corresponding local context region. This step effectively updates the coarse point query, transforming it into an enriched one for the subsequent LLM reasoning. Extensive experiments demonstrate that our approach compresses the visual tokens by 75%~89%, while achieves comparable or even better performance across diverse benchmarks with significantly higher efficiency. The source codes can be found at https://github.com/CircleRadon/TokenPacker.
Authors: Shuohao Shi, Qiang Fang, Tong Zhao, Xin Xu
Abstract: Tiny object detection is becoming one of the most challenging tasks in computer vision because of the limited object size and lack of information. The label assignment strategy is a key factor affecting the accuracy of object detection. Although there are some effective label assignment strategies for tiny objects, most of them focus on reducing the sensitivity to the bounding boxes to increase the number of positive samples and have some fixed hyperparameters need to set. However, more positive samples may not necessarily lead to better detection results, in fact, excessive positive samples may lead to more false positives. In this paper, we introduce a simple but effective strategy named the Similarity Distance (SimD) to evaluate the similarity between bounding boxes. This proposed strategy not only considers both location and shape similarity but also learns hyperparameters adaptively, ensuring that it can adapt to different datasets and various object sizes in a dataset. Our approach can be simply applied in common anchor-based detectors in place of the IoU for label assignment and Non Maximum Suppression (NMS). Extensive experiments on four mainstream tiny object detection datasets demonstrate superior performance of our method, especially, 1.8 AP points and 4.1 AP points of very tiny higher than the state-of-the-art competitors on AI-TOD. Code is available at: \url{https://github.com/cszzshi/SimD}.
Authors: Ling Yang, Zixiang Zhang, Zhilong Zhang, Xingchao Liu, Minkai Xu, Wentao Zhang, Chenlin Meng, Stefano Ermon, Bin Cui
Abstract: Flow matching (FM) is a general framework for defining probability paths via Ordinary Differential Equations (ODEs) to transform between noise and data samples. Recent approaches attempt to straighten these flow trajectories to generate high-quality samples with fewer function evaluations, typically through iterative rectification methods or optimal transport solutions. In this paper, we introduce Consistency Flow Matching (Consistency-FM), a novel FM method that explicitly enforces self-consistency in the velocity field. Consistency-FM directly defines straight flows starting from different times to the same endpoint, imposing constraints on their velocity values. Additionally, we propose a multi-segment training approach for Consistency-FM to enhance expressiveness, achieving a better trade-off between sampling quality and speed. Preliminary experiments demonstrate that our Consistency-FM significantly improves training efficiency by converging 4.4x faster than consistency models and 1.7x faster than rectified flow models while achieving better generation quality. Our code is available at: https://github.com/YangLing0818/consistency_flow_matching
URLs: https://github.com/YangLing0818/consistency_flow_matching
Authors: Yoon Gyo Jung, Jaewoo Park, Xingbo Dong, Hojin Park, Andrew Beng Jin Teoh, Octavia Camps
Abstract: Understanding the vulnerability of face recognition systems to malicious attacks is of critical importance. Previous works have focused on reconstructing face images that can penetrate a targeted verification system. Even in the white-box scenario, however, naively reconstructed images misrepresent the identity information, hence the attacks are easily neutralized once the face system is updated or changed. In this paper, we aim to reconstruct face images which are capable of transferring face attacks on unseen encoders. We term this problem as Face Reconstruction Transfer Attack (FRTA) and show that it can be formulated as an out-of-distribution (OOD) generalization problem. Inspired by its OOD nature, we propose to solve FRTA by Averaged Latent Search and Unsupervised Validation with pseudo target (ALSUV). To strengthen the reconstruction attack on OOD unseen encoders, ALSUV reconstructs the face by searching the latent of amortized generator StyleGAN2 through multiple latent optimization, latent optimization trajectory averaging, and unsupervised validation with a pseudo target. We demonstrate the efficacy and generalization of our method on widely used face datasets, accompanying it with extensive ablation studies and visually, qualitatively, and quantitatively analyses. The source code will be released.
Authors: Jinmin Li, Kuofeng Gao, Yang Bai, Jingyun Zhang, Shu-Tao Xia
Abstract: The advent of video-based Large Language Models (LLMs) has significantly enhanced video understanding. However, it has also raised some safety concerns regarding data protection, as videos can be more easily annotated, even without authorization. This paper introduces Video Watermarking, a novel technique to protect videos from unauthorized annotations by such video-based LLMs, especially concerning the video content and description, in response to specific queries. By imperceptibly embedding watermarks into key video frames with multi-modal flow-based losses, our method preserves the viewing experience while preventing misuse by video-based LLMs. Extensive experiments show that Video Watermarking significantly reduces the comprehensibility of videos with various video-based LLMs, demonstrating both stealth and robustness. In essence, our method provides a solution for securing video content, ensuring its integrity and confidentiality in the face of evolving video-based LLMs technologies.
Authors: Sergio Izquierdo, Javier Civera
Abstract: Visual Place Recognition (VPR) plays a critical role in many localization and mapping pipelines. It consists of retrieving the closest sample to a query image, in a certain embedding space, from a database of geotagged references. The image embedding is learned to effectively describe a place despite variations in visual appearance, viewpoint, and geometric changes. In this work, we formulate how limitations in the Geographic Distance Sensitivity of current VPR embeddings result in a high probability of incorrectly sorting the top-k retrievals, negatively impacting the recall. In order to address this issue in single-stage VPR, we propose a novel mining strategy, CliqueMining, that selects positive and negative examples by sampling cliques from a graph of visually similar images. Our approach boosts the sensitivity of VPR embeddings at small distance ranges, significantly improving the state of the art on relevant benchmarks. In particular, we raise recall@1 from 75% to 82% in MSLS Challenge, and from 76% to 90% in Nordland. Models and code are available at https://github.com/serizba/cliquemining.
Authors: Raphael Bensadoun, Yanir Kleiman, Idan Azuri, Omri Harosh, Andrea Vedaldi, Natalia Neverova, Oran Gafni
Abstract: The recent availability and adaptability of text-to-image models has sparked a new era in many related domains that benefit from the learned text priors as well as high-quality and fast generation capabilities, one of which is texture generation for 3D objects. Although recent texture generation methods achieve impressive results by using text-to-image networks, the combination of global consistency, quality, and speed, which is crucial for advancing texture generation to real-world applications, remains elusive. To that end, we introduce Meta 3D TextureGen: a new feedforward method comprised of two sequential networks aimed at generating high-quality and globally consistent textures for arbitrary geometries of any complexity degree in less than 20 seconds. Our method achieves state-of-the-art results in quality and speed by conditioning a text-to-image model on 3D semantics in 2D space and fusing them into a complete and high-resolution UV texture map, as demonstrated by extensive qualitative and quantitative evaluations. In addition, we introduce a texture enhancement network that is capable of up-scaling any texture by an arbitrary ratio, producing 4k pixel resolution textures.
Authors: Souradeep Chakraborty, Zijun Wei, Conor Kelton, Seoyoung Ahn, Aruna Balasubramanian, Gregory J. Zelinsky, Dimitris Samaras
Abstract: We present a model for predicting visual attention during the free viewing of graphic design documents. While existing works on this topic have aimed at predicting static saliency of graphic designs, our work is the first attempt to predict both spatial attention and dynamic temporal order in which the document regions are fixated by gaze using a deep learning based model. We propose a two-stage model for predicting dynamic attention on such documents, with webpages being our primary choice of document design for demonstration. In the first stage, we predict the saliency maps for each of the document components (e.g. logos, banners, texts, etc. for webpages) conditioned on the type of document layout. These component saliency maps are then jointly used to predict the overall document saliency. In the second stage, we use these layout-specific component saliency maps as the state representation for an inverse reinforcement learning model of fixation scanpath prediction during document viewing. To test our model, we collected a new dataset consisting of eye movements from 41 people freely viewing 450 webpages (the largest dataset of its kind). Experimental results show that our model outperforms existing models in both saliency and scanpath prediction for webpages, and also generalizes very well to other graphic design documents such as comics, posters, mobile UIs, etc. and natural images.
Authors: Yawar Siddiqui, Tom Monnier, Filippos Kokkinos, Mahendra Kariya, Yanir Kleiman, Emilien Garreau, Oran Gafni, Natalia Neverova, Andrea Vedaldi, Roman Shapovalov, David Novotny
Abstract: We present Meta 3D AssetGen (AssetGen), a significant advancement in text-to-3D generation which produces faithful, high-quality meshes with texture and material control. Compared to works that bake shading in the 3D object's appearance, AssetGen outputs physically-based rendering (PBR) materials, supporting realistic relighting. AssetGen generates first several views of the object with factored shaded and albedo appearance channels, and then reconstructs colours, metalness and roughness in 3D, using a deferred shading loss for efficient supervision. It also uses a sign-distance function to represent 3D shape more reliably and introduces a corresponding loss for direct shape supervision. This is implemented using fused kernels for high memory efficiency. After mesh extraction, a texture refinement transformer operating in UV space significantly improves sharpness and details. AssetGen achieves 17% improvement in Chamfer Distance and 40% in LPIPS over the best concurrent work for few-view reconstruction, and a human preference of 72% over the best industry competitors of comparable speed, including those that support PBR. Project page with generated assets: https://assetgen.github.io
Authors: Bo Zhang, Zimeng Zhou, Boyu Jiang, Rong Zheng
Abstract: In industrial countries, adults spend a considerable amount of time sedentary each day at work, driving and during activities of daily living. Characterizing the seated upper body human poses using mmWave radars is an important, yet under-studied topic with many applications in human-machine interaction, transportation and road safety. In this work, we devise SUPER, a framework for seated upper body human pose estimation that utilizes dual-mmWave radars in close proximity. A novel masking algorithm is proposed to coherently fuse data from the radars to generate intensity and Doppler point clouds with complementary information for high-motion but small radar cross section areas (e.g., upper extremities) and low-motion but large RCS areas (e.g. torso). A lightweight neural network extracts both global and local features of upper body and output pose parameters for the Skinned Multi-Person Linear (SMPL) model. Extensive leave-one-subject-out experiments on various motion sequences from multiple subjects show that SUPER outperforms a state-of-the-art baseline method by 30 -- 184%. We also demonstrate its utility in a simple downstream task for hand-object interaction.
Authors: Elmira Amirloo, Jean-Philippe Fauconnier, Christoph Roesmann, Christian Kerl, Rinu Boney, Yusu Qian, Zirui Wang, Afshin Dehghan, Yinfei Yang, Zhe Gan, Peter Grasch
Abstract: Preference alignment has become a crucial component in enhancing the performance of Large Language Models (LLMs), yet its impact in Multimodal Large Language Models (MLLMs) remains comparatively underexplored. Similar to language models, MLLMs for image understanding tasks encounter challenges like hallucination. In MLLMs, hallucination can occur not only by stating incorrect facts but also by producing responses that are inconsistent with the image content. A primary objective of alignment for MLLMs is to encourage these models to align responses more closely with image information. Recently, multiple works have introduced preference datasets for MLLMs and examined different alignment methods, including Direct Preference Optimization (DPO) and Proximal Policy Optimization (PPO). However, due to variations in datasets, base model types, and alignment methods, it remains unclear which specific elements contribute most significantly to the reported improvements in these works. In this paper, we independently analyze each aspect of preference alignment in MLLMs. We start by categorizing the alignment algorithms into two groups, offline (such as DPO), and online (such as online-DPO), and show that combining offline and online methods can improve the performance of the model in certain scenarios. We review a variety of published multimodal preference datasets and discuss how the details of their construction impact model performance. Based on these insights, we introduce a novel way of creating multimodal preference data called Bias-Driven Hallucination Sampling (BDHS) that needs neither additional annotation nor external models, and show that it can achieve competitive performance to previously published alignment work for multimodal models across a range of benchmarks.
Authors: Fei Shen, Hu Ye, Sibo Liu, Jun Zhang, Cong Wang, Xiao Han, Wei Yang
Abstract: Recent research showcases the considerable potential of conditional diffusion models for generating consistent stories. However, current methods, which predominantly generate stories in an autoregressive and excessively caption-dependent manner, often underrate the contextual consistency and relevance of frames during sequential generation. To address this, we propose a novel Rich-contextual Conditional Diffusion Models (RCDMs), a two-stage approach designed to enhance story generation's semantic consistency and temporal consistency. Specifically, in the first stage, the frame-prior transformer diffusion model is presented to predict the frame semantic embedding of the unknown clip by aligning the semantic correlations between the captions and frames of the known clip. The second stage establishes a robust model with rich contextual conditions, including reference images of the known clip, the predicted frame semantic embedding of the unknown clip, and text embeddings of all captions. By jointly injecting these rich contextual conditions at the image and feature levels, RCDMs can generate semantic and temporal consistency stories. Moreover, RCDMs can generate consistent stories with a single forward inference compared to autoregressive models. Our qualitative and quantitative results demonstrate that our proposed RCDMs outperform in challenging scenarios. The code and model will be available at https://github.com/muzishen/RCDMs.
Authors: Nataniel Ruiz, Yuanzhen Li, Neal Wadhwa, Yael Pritch, Michael Rubinstein, David E. Jacobs, Shlomi Fruchter
Abstract: We present Magic Insert, a method for dragging-and-dropping subjects from a user-provided image into a target image of a different style in a physically plausible manner while matching the style of the target image. This work formalizes the problem of style-aware drag-and-drop and presents a method for tackling it by addressing two sub-problems: style-aware personalization and realistic object insertion in stylized images. For style-aware personalization, our method first fine-tunes a pretrained text-to-image diffusion model using LoRA and learned text tokens on the subject image, and then infuses it with a CLIP representation of the target style. For object insertion, we use Bootstrapped Domain Adaption to adapt a domain-specific photorealistic object insertion model to the domain of diverse artistic styles. Overall, the method significantly outperforms traditional approaches such as inpainting. Finally, we present a dataset, SubjectPlop, to facilitate evaluation and future progress in this area. Project page: https://magicinsert.github.io/
Authors: Dohyun Kim, Nayoung Oh, Deokmin Hwang, Daehyung Park
Abstract: We aim to solve the problem of spatially localizing composite instructions referring to space: space grounding. Compared to current instance grounding, space grounding is challenging due to the ill-posedness of identifying locations referred to by discrete expressions and the compositional ambiguity of referring expressions. Therefore, we propose a novel probabilistic space-grounding methodology (LINGO-Space) that accurately identifies a probabilistic distribution of space being referred to and incrementally updates it, given subsequent referring expressions leveraging configurable polar distributions. Our evaluations show that the estimation using polar distributions enables a robot to ground locations successfully through $20$ table-top manipulation benchmark tests. We also show that updating the distribution helps the grounding method accurately narrow the referring space. We finally demonstrate the robustness of the space grounding with simulated manipulation and real quadruped robot navigation tasks. Code and videos are available at https://lingo-space.github.io.
Authors: Shijian Ma, Shicong Ma, Weize Ma
Abstract: This paper presents WaveShot, an innovative portable unmanned surface vessel that aims to transform water surface videography by offering a highly maneuverable, cost-effective, and safe alternative to traditional filming methods. WaveShot is specially designed for the modern demands of film production, advertising, documentaries, and visual arts, equipped with professional-grade waterproof cameras and advanced technology to capture both static and dynamic scenes on waterways. We discuss the development and advantages of WaveShot, highlighting its portability, ease of transport, and rapid deployment capabilities. Experimental validation that is showcasing WaveShot's stability and high-quality video capture in various water conditions, and the integration of monocular depth estimation algorithms to enhance the operator's spatial perception. The paper concludes with an exploration of WaveShot's real-world applications, its user-friendly remote operation, and future enhancements such as gimbal integration and advanced computer vision for optimized videography on water surfaces.
Authors: Alexander Denker, Zeljko Kereta, Imraj Singh, Tom Freudenberg, Tobias Kluth, Peter Maass, Simon Arridge
Abstract: Electrical impedance tomography (EIT) plays a crucial role in non-invasive imaging, with both medical and industrial applications. In this paper, we present three data-driven reconstruction methods for EIT imaging. These three approaches were originally submitted to the Kuopio tomography challenge 2023 (KTC2023). First, we introduce a post-processing approach, which achieved first place at KTC2023. Further, we present a fully learned and a conditional diffusion approach. All three methods are based on a similar neural network as a backbone and were trained using a synthetically generated data set, providing with an opportunity for a fair comparison of these different data-driven reconstruction methods.
Authors: Elena Sizikova, Andreu Badal, Jana G. Delfino, Miguel Lago, Brandon Nelson, Niloufar Saharkhiz, Berkman Sahiner, Ghada Zamzmi, Aldo Badano
Abstract: A key challenge for the development and deployment of artificial intelligence (AI) solutions in radiology is solving the associated data limitations. Obtaining sufficient and representative patient datasets with appropriate annotations may be burdensome due to high acquisition cost, safety limitations, patient privacy restrictions or low disease prevalence rates. In silico data offers a number of potential advantages to patient data, such as diminished patient harm, reduced cost, simplified data acquisition, scalability, improved quality assurance testing, and a mitigation approach to data imbalances. We summarize key research trends and practical uses for synthetically generated data for radiological applications of AI. Specifically, we discuss different types of techniques for generating synthetic examples, their main application areas, and related quality control assessment issues. We also discuss current approaches for evaluating synthetic imaging data. Overall, synthetic data holds great promise in addressing current data availability gaps, but additional work is needed before its full potential is realized.
Authors: Haibo Jin, Leyang Hu, Xinuo Li, Peiyan Zhang, Chonghan Chen, Jun Zhuang, Haohan Wang
Abstract: The rapid evolution of artificial intelligence (AI) through developments in Large Language Models (LLMs) and Vision-Language Models (VLMs) has brought significant advancements across various technological domains. While these models enhance capabilities in natural language processing and visual interactive tasks, their growing adoption raises critical concerns regarding security and ethical alignment. This survey provides an extensive review of the emerging field of jailbreaking--deliberately circumventing the ethical and operational boundaries of LLMs and VLMs--and the consequent development of defense mechanisms. Our study categorizes jailbreaks into seven distinct types and elaborates on defense strategies that address these vulnerabilities. Through this comprehensive examination, we identify research gaps and propose directions for future studies to enhance the security frameworks of LLMs and VLMs. Our findings underscore the necessity for a unified perspective that integrates both jailbreak strategies and defensive solutions to foster a robust, secure, and reliable environment for the next generation of language models. More details can be found on our website: \url{https://chonghan-chen.com/llm-jailbreak-zoo-survey/}.
Authors: Ruochen Wang, Ting Liu, Cho-Jui Hsieh, Boqing Gong
Abstract: This paper introduces the first gradient-based framework for prompt optimization in text-to-image diffusion models. We formulate prompt engineering as a discrete optimization problem over the language space. Two major challenges arise in efficiently finding a solution to this problem: (1) Enormous Domain Space: Setting the domain to the entire language space poses significant difficulty to the optimization process. (2) Text Gradient: Efficiently computing the text gradient is challenging, as it requires backpropagating through the inference steps of the diffusion model and a non-differentiable embedding lookup table. Beyond the problem formulation, our main technical contributions lie in solving the above challenges. First, we design a family of dynamically generated compact subspaces comprised of only the most relevant words to user input, substantially restricting the domain space. Second, we introduce "Shortcut Text Gradient" -- an effective replacement for the text gradient that can be obtained with constant memory and runtime. Empirical evaluation on prompts collected from diverse sources (DiffusionDB, ChatGPT, COCO) suggests that our method can discover prompts that substantially improve (prompt enhancement) or destroy (adversarial attack) the faithfulness of images generated by the text-to-image diffusion model.
Authors: Jingxi Chen, Botao He, Chahat Deep Singh, Cornelia Fermuller, Yiannis Aloimonos
Abstract: One of the core activities of an active observer involves moving to secure a "better" view of the scene, where the definition of "better" is task-dependent. This paper focuses on the task of human pose estimation from videos capturing a person's activity. Self-occlusions within the scene can complicate or even prevent accurate human pose estimation. To address this, relocating the camera to a new vantage point is necessary to clarify the view, thereby improving 2D human pose estimation. This paper formalizes the process of achieving an improved viewpoint. Our proposed solution to this challenge comprises three main components: a NeRF-based Drone-View Data Generation Framework, an On-Drone Network for Camera View Error Estimation, and a Combined Planner for devising a feasible motion plan to reposition the camera based on the predicted errors for camera views. The Data Generation Framework utilizes NeRF-based methods to generate a comprehensive dataset of human poses and activities, enhancing the drone's adaptability in various scenarios. The Camera View Error Estimation Network is designed to evaluate the current human pose and identify the most promising next viewing angles for the drone, ensuring a reliable and precise pose estimation from those angles. Finally, the combined planner incorporates these angles while considering the drone's physical and environmental limitations, employing efficient algorithms to navigate safe and effective flight paths. This system represents a significant advancement in active 2D human pose estimation for an autonomous UAV agent, offering substantial potential for applications in aerial cinematography by improving the performance of autonomous human pose estimation and maintaining the operational safety and efficiency of UAVs.
Authors: Wenyi Lian, Joakim Lindblad, Christina Runow Stark, Jan-Micha\'el Hirsch, Nata\v{s}a Sladoje
Abstract: Oral cancer is a global health challenge. It is treatable if detected early, but it is often fatal in late stages. There is a shift from the invasive and time-consuming tissue sampling and histological examination, toward non-invasive brush biopsies and cytological examination. Reliable computer-assisted methods are essential for cost-effective and accurate cytological analysis, but the lack of detailed cell-level annotations impairs model effectiveness. This study aims to improve AI-based oral cancer detection using multimodal imaging and deep fusion. We combine brightfield and fluorescence whole slide microscopy imaging to analyze Papanicolaou-stained liquid-based cytology slides of brush biopsies collected from both healthy and cancer patients. Due to limited cytological annotations, we utilize a weakly supervised deep learning approach using only patient-level labels. We evaluate various multimodal fusion strategies, including early, late, and three recent intermediate fusion methods. Our results show: (i) fluorescence imaging of Papanicolaou-stained samples provides substantial diagnostic information; (ii) multimodal fusion enhances classification and cancer detection accuracy over single-modality methods. Intermediate fusion is the leading method among the studied approaches. Specifically, the Co-Attention Fusion Network (CAFNet) model excels with an F1 score of 83.34% and accuracy of 91.79%, surpassing human performance on the task. Additional tests highlight the need for precise image registration to optimize multimodal analysis benefits. This study advances cytopathology by combining deep learning and multimodal imaging to enhance early, non-invasive detection of oral cancer, improving diagnostic accuracy and streamlining clinical workflows. The developed pipeline is also applicable in other cytological settings. Our codes and dataset are available online for further research.
Authors: Du Xueyu, Zhang Lilian, Liu Ruochen, Wang Maosong, Wu Wenqi, Mao Jun
Abstract: Efficient Visual-Inertial Odometry (VIO) is crucial for payload-constrained robots. Though modern optimization-based algorithms have achieved superior accuracy, the MSCKF-based VIO algorithms are still widely demanded for their efficient and consistent performance. As MSCKF is built upon the conventional multi-view geometry, the measured residuals are not only related to the state errors but also related to the feature position errors. To apply EKF fusion, a projection process is required to remove the feature position error from the observation model, which can lead to model and accuracy degradation. To obtain an efficient visual-inertial fusion model, while also preserving the model consistency, we propose to reconstruct the MSCKF VIO with the novel Pose-Only (PO) multi-view geometry description. In the newly constructed filter, we have modeled PO reprojection residuals, which are solely related to the motion states and thus overcome the requirements of space projection. Moreover, the new filter does not require any feature position information, which removes the computational cost and linearization errors brought in by the 3D reconstruction procedure. We have conducted comprehensive experiments on multiple datasets, where the proposed method has shown accuracy improvements and consistent performance in challenging sequences.
Authors: Calvin Luo, Mandy He, Zilai Zeng, Chen Sun
Abstract: Training an agent to achieve particular goals or perform desired behaviors is often accomplished through reinforcement learning, especially in the absence of expert demonstrations. However, supporting novel goals or behaviors through reinforcement learning requires the ad-hoc design of appropriate reward functions, which quickly becomes intractable. To address this challenge, we propose Text-Aware Diffusion for Policy Learning (TADPoLe), which uses a pretrained, frozen text-conditioned diffusion model to compute dense zero-shot reward signals for text-aligned policy learning. We hypothesize that large-scale pretrained generative models encode rich priors that can supervise a policy to behave not only in a text-aligned manner, but also in alignment with a notion of naturalness summarized from internet-scale training data. In our experiments, we demonstrate that TADPoLe is able to learn policies for novel goal-achievement and continuous locomotion behaviors specified by natural language, in both Humanoid and Dog environments. The behaviors are learned zero-shot without ground-truth rewards or expert demonstrations, and are qualitatively more natural according to human evaluation. We further show that TADPoLe performs competitively when applied to robotic manipulation tasks in the Meta-World environment.
Authors: Tongtong Zhang, Zongcheng Zuo, Yuanxiang Li
Abstract: Digital Elevation Model (DEM) plays a fundamental role in remote sensing and photogrammetry. Enhancing the quality of DEM is crucial for various applications. Although multiple types of defects may appear simultaneously in the same DEM, they are commonly addressed separately. Most existing approaches only aim to fill the DEM voids, or apply super-resolution to the intact DEM. This paper introduces a unified generative model that simultaneously addresses voids and low-resolution problems, rather than taking two separate measures. The proposed approach presents the DEM Stochastic Differential Equation (DEM-SDE) for unified DEM quality enhancement. The DEM degradation of downsampling and random voids adding is modeled as the SDE forwarding, and the restoration is achieved by simulating the corresponding revert process. Conditioned on the terrain feature, and adopting efficient submodules with lightweight channel attention, DEM-SDE simultaneously enhances the DEM quality with an efficient process for training. The experiments show that DEM-SDE method achieves highly competitive performance in simultaneous super-resolution and void filling compared to the state-of-the-art work. DEM-SDE also manifests robustness for larger DEM patches.
Authors: Zitao Chen, Karthik Pattabiraman
Abstract: Modern machine learning (ML) ecosystems offer a surging number of ML frameworks and code repositories that can greatly facilitate the development of ML models. Today, even ordinary data holders who are not ML experts can apply off-the-shelf codebase to build high-performance ML models on their data, many of which are sensitive in nature (e.g., clinical records). In this work, we consider a malicious ML provider who supplies model-training code to the data holders, does not have access to the training process, and has only black-box query access to the resulting model. In this setting, we demonstrate a new form of membership inference attack that is strictly more powerful than prior art. Our attack empowers the adversary to reliably de-identify all the training samples (average >99% attack TPR@0.1% FPR), and the compromised models still maintain competitive performance as their uncorrupted counterparts (average <1% accuracy drop). Moreover, we show that the poisoned models can effectively disguise the amplified membership leakage under common membership privacy auditing, which can only be revealed by a set of secret samples known by the adversary. Overall, our study not only points to the worst-case membership privacy leakage, but also unveils a common pitfall underlying existing privacy auditing methods, which calls for future efforts to rethink the current practice of auditing membership privacy in machine learning models.
Authors: Bozhong Tian, Xiaozhuan Liang, Siyuan Cheng, Qingbin Liu, Mengru Wang, Dianbo Sui, Xi Chen, Huajun Chen, Ningyu Zhang
Abstract: Large Language Models (LLMs) trained on extensive corpora inevitably retain sensitive data, such as personal privacy information and copyrighted material. Recent advancements in knowledge unlearning involve updating LLM parameters to erase specific knowledge. However, current unlearning paradigms are mired in vague forgetting boundaries, often erasing knowledge indiscriminately. In this work, we introduce KnowUnDo, a benchmark containing copyrighted content and user privacy domains to evaluate if the unlearning process inadvertently erases essential knowledge. Our findings indicate that existing unlearning methods often suffer from excessive unlearning. To address this, we propose a simple yet effective method, MemFlex, which utilizes gradient information to precisely target and unlearn sensitive parameters. Experimental results show that MemFlex is superior to existing methods in both precise knowledge unlearning and general knowledge retaining of LLMs. Code and dataset will be released at https://github.com/zjunlp/KnowUnDo.
Authors: Chaoxing Huang, Ziqiang Yu, Zijian Gao, Qiuyi Shen, Queenie Chan, Vincent Wai-Sun Wong, Winnie Chiu-Wing Chu, Weitian Chen
Abstract: This study evaluated a deep learning-based method using Deep Image Prior (DIP) to quantify triglyceride double bonds from chemical-shift encoded multi-echo gradient echo images without network training. We employed a cost function based on signal constraints to iteratively update the neural network on a single dataset. The method was validated using phantom experiments and in vivo scans. Results showed close alignment between measured and reference double bond values, with phantom experiments yielding a Pearson correlation coefficient of 0.96 (p = .0005). In vivo results demonstrated good agreement in subcutaneous fat. We conclude that Deep Image Prior shows feasibility for quantifying double bonds and fatty acid content from chemical-shift encoded multi-echo MRI.
Authors: Khyathi Raghavi Chandu, Linjie Li, Anas Awadalla, Ximing Lu, Jae Sung Park, Jack Hessel, Lijuan Wang, Yejin Choi
Abstract: The ability to acknowledge the inevitable uncertainty in their knowledge and reasoning is a prerequisite for AI systems to be truly truthful and reliable. In this paper, we present a taxonomy of uncertainty specific to vision-language AI systems, distinguishing between epistemic uncertainty (arising from a lack of information) and aleatoric uncertainty (due to inherent unpredictability), and further explore finer categories within. Based on this taxonomy, we synthesize a benchmark dataset, CertainlyUncertain, featuring 178K visual question answering (VQA) samples as contrastive pairs. This is achieved by 1) inpainting images to make previously answerable questions into unanswerable ones; and 2) using image captions to prompt large language models for both answerable and unanswerable questions. Additionally, we introduce a new metric confidence-weighted accuracy, that is well correlated with both accuracy and calibration error, to address the shortcomings of existing metrics.
Authors: Dennis Mronga, Andreas Bresser, Fabian Maas, Adrian Danzglock, Simon Stelter, Alina Hawkin, Hoang Giang Nguyen, Michael Beetz, Frank Kirchner
Abstract: In this paper, we present the service robot MARLIN and its integration with the K4R platform, a cloud system for complex AI applications in retail. At its core, this platform contains so-called semantic digital twins, a semantically annotated representation of the retail store. MARLIN continuously exchanges data with the K4R platform, improving the robot's capabilities in perception, autonomous navigation, and task planning. We exploit these capabilities in a retail intralogistics scenario, specifically by assisting store employees in stocking shelves. We demonstrate that MARLIN is able to update the digital representation of the retail store by detecting and classifying obstacles, autonomously planning and executing replenishment missions, adapting to unforeseen changes in the environment, and interacting with store employees. Experiments are conducted in simulation, in a laboratory environment, and in a real store. We also describe and evaluate a novel algorithm for autonomous navigation of articulated tractor-trailer systems. The algorithm outperforms the manufacturer's proprietary navigation approach and improves MARLIN's navigation capabilities in confined spaces.
Authors: Paolo Panicucci, Fabio Ornati, Francesco Topputo
Abstract: The increasing interest in spacecraft autonomy and the complex tasks to be accomplished by the spacecraft raise the need for a trustworthy approach to perform Verification & Validation of Guidance, Navigation, and Control algorithms. In the context of autonomous operations, vision-based navigation algorithms have established themselves as effective solutions to determine the spacecraft state in orbit with low-cost and versatile sensors. Nevertheless, detailed testing must be performed on ground to understand the algorithm's robustness and performance on flight hardware. Given the impossibility of testing directly on orbit these algorithms, a dedicated simulation framework must be developed to emulate the orbital environment in a laboratory setup. This paper presents the design of a low-aberration optical facility called RETINA to perform this task. RETINA is designed to accommodate cameras with different characteristics (e.g., sensor size and focal length) while ensuring the correct stimulation of the camera detector. A preliminary design is performed to identify the range of possible components to be used in the facility according to the facility requirements. Then, a detailed optical design is performed in Zemax OpticStudio to optimize the number and characteristics of the lenses composing the facility's optical systems. The final design is compared against the preliminary design to show the superiority of the optical performance achieved with this approach. This work presents also a calibration procedure to estimate the misalignment and the centering errors in the facility. These estimated parameters are used in a dedicated compensation algorithm, enabling the stimulation of the camera at tens of arcseconds of precision. Finally, two different applications are presented to show the versatility of RETINA in accommodating different cameras and in simulating different mission scenarios.
Authors: Shuaike Xu, Xiaolin Zhang, Peng Zhang, Kun Zhan
Abstract: Graph node classification with few labeled nodes presents significant challenges due to limited supervision. Conventional methods often exploit the graph in a transductive learning manner. They fail to effectively utilize the abundant unlabeled data and the structural information inherent in graphs. To address these issues, we introduce a Structure-Aware Consensus Network (SACN) from three perspectives. Firstly, SACN leverages a novel structure-aware consensus learning strategy between two strongly augmented views. The proposed strategy can fully exploit the potentially useful information of the unlabeled nodes and the structural information of the entire graph. Secondly, SACN uniquely integrates the graph's structural information to achieve strong-to-strong consensus learning, improving the utilization of unlabeled data while maintaining multiview learning. Thirdly, unlike two-branch graph neural network-based methods, SACN is designed for multiview feature learning within a single-branch architecture. Furthermore, a class-aware pseudolabel selection strategy helps address class imbalance and achieve effective weak-to-strong supervision. Extensive experiments on three benchmark datasets demonstrate SACN's superior performance in node classification tasks, particularly at very low label rates, outperforming state-of-the-art methods while maintaining computational simplicity.The source code is available at https://github.com/kunzhan/SACN
Authors: JiaQi Luo
Abstract: Against the backdrop of advancing science and technology, autonomous vehicle technology has emerged as a focal point of intense scrutiny within the academic community. Nevertheless, the challenge persists in guaranteeing the safety and reliability of this technology when navigating intricate scenarios. While a substantial portion of autonomous driving research is dedicated to testing in open-air environments, such as urban roads and highways, where the myriad variables at play are meticulously examined, enclosed indoor spaces like underground parking lots have, to a significant extent, been overlooked in the scholarly discourse. This discrepancy highlights a gap in derstanding the unique challenges these confined settings pose for autonomous navigation systems. This study tackles indoor autonomous driving, particularly in overlooked spaces like underground parking lots. Using CARLA's simulation platform, a realistic parking model is created for data gathering. An occupancy grid network then processes this data to predict vehicle paths and obstacles, enhancing the system's perception in complex indoor environments. Ultimately, this strategy improves safety in autonomous parking operations. The paper meticulously evaluates the model's predictive capabilities, validating its efficacy in the context of underground parking. Our findings confirm that the proposed strategy successfully enhances autonomous vehicle performance in these complex indoor settings. It equips autonomous systems with improved adaptation to underground lots, reinforcing safety measures and dependability. This work paves the way for future advancements and applications by addressing the research shortfall concerning indoor parking environments, serving as a pivotal reference point.
Authors: Jiaxu Tian, Fan Lyu
Abstract: Continual Test-Time Adaptation (CTTA) aims to adapt a pretrained model to ever-changing environments during the test time under continuous domain shifts. Most existing CTTA approaches are based on the Mean Teacher (MT) structure, which contains a student and a teacher model, where the student is updated using the pseudo-labels from the teacher model, and the teacher is then updated by exponential moving average strategy. However, these methods update the MT model indiscriminately on all parameters of the model. That is, some critical parameters involving sharing knowledge across different domains may be erased, intensifying error accumulation and catastrophic forgetting. In this paper, we introduce Parameter-Selective Mean Teacher (PSMT) method, which is capable of effectively updating the critical parameters within the MT network under domain shifts. First, we introduce a selective distillation mechanism in the student model, which utilizes past knowledge to regularize novel knowledge, thereby mitigating the impact of error accumulation. Second, to avoid catastrophic forgetting, in the teacher model, we create a mask through Fisher information to selectively update parameters via exponential moving average, with preservation measures applied to crucial parameters. Extensive experimental results verify that PSMT outperforms state-of-the-art methods across multiple benchmark datasets. Our code is available at \url{https://github.com/JiaxuTian/PSMT}.
Authors: Yurui Huang, Yang Yang, Shou Chen, Xiangyu Wu, Qingguo Chen, Jianfeng Lu
Abstract: In this paper, we propose a solution for improving the quality of temporal sound localization. We employ a multimodal fusion approach to combine visual and audio features. High-quality visual features are extracted using a state-of-the-art self-supervised pre-training network, resulting in efficient video feature representations. At the same time, audio features serve as complementary information to help the model better localize the start and end of sounds. The fused features are trained in a multi-scale Transformer for training. In the final test dataset, we achieved a mean average precision (mAP) of 0.33, obtaining the second-best performance in this track.
Authors: Musashi Hinck, Carolin Holtermann, Matthew Lyle Olson, Florian Schneider, Sungduk Yu, Anahita Bhiwandiwalla, Anne Lauscher, Shaoyen Tseng, Vasudev Lal
Abstract: We uncover a surprising multilingual bias occurring in a popular class of multimodal vision-language models (VLMs). Including an image in the query to a LLaVA-style VLM significantly increases the likelihood of the model returning an English response, regardless of the language of the query. This paper investigates the causes of this loss with a two-pronged approach that combines extensive ablation of the design space with a mechanistic analysis of the models' internal representations of image and text inputs. Both approaches indicate that the issue stems in the language modelling component of the LLaVA model. Statistically, we find that switching the language backbone for a bilingual language model has the strongest effect on reducing this error. Mechanistically, we provide compelling evidence that visual inputs are not mapped to a similar space as text ones, and that intervening on intermediary attention layers can reduce this bias. Our findings provide important insights to researchers and engineers seeking to understand the crossover between multimodal and multilingual spaces, and contribute to the goal of developing capable and inclusive VLMs for non-English contexts.
Authors: Lorenzo S. Querol, Hajime Nagahara, Hideaki Hayashi
Abstract: The growing use of deep learning in safety-critical applications, such as medical imaging, has raised concerns about limited labeled data, where this demand is amplified as model complexity increases, posing hurdles for domain experts to annotate data. In response to this, active learning (AL) is used to efficiently train models with limited annotation costs. In the context of deep neural networks (DNNs), AL often uses confidence or probability outputs as a score for selecting the most informative samples. However, modern DNNs exhibit unreliable confidence outputs, making calibration essential. We propose an AL framework that self-calibrates the confidence used for sample selection during the training process, referred to as Confident Active Learning with Integrated CalibratiOn (CALICO). CALICO incorporates the joint training of a classifier and an energy-based model, instead of the standard softmax-based classifier. This approach allows for simultaneous estimation of the input data distribution and the class probabilities during training, improving calibration without needing an additional labeled dataset. Experimental results showcase improved classification performance compared to a softmax-based classifier with fewer labeled samples. Furthermore, the calibration stability of the model is observed to depend on the prior class distribution of the data.
Authors: Zhipeng Deng, Luyang Luo, Hao Chen
Abstract: The right to be forgotten, as stated in most data regulations, poses an underexplored challenge in federated learning (FL), leading to the development of federated unlearning (FU). However, current FU approaches often face trade-offs between efficiency, model performance, forgetting efficacy, and privacy preservation. In this paper, we delve into the paradigm of Federated Client Unlearning (FCU) to guarantee a client the right to erase the contribution or the influence, introducing the first FU framework in medical imaging. In the unlearning process of a client, the proposed model-contrastive unlearning marks a pioneering step towards feature-level unlearning, and frequency-guided memory preservation ensures smooth forgetting of local knowledge while maintaining the generalizability of the trained global model, thus avoiding performance compromises and guaranteeing rapid post-training. We evaluated our FCU framework on two public medical image datasets, including Intracranial hemorrhage diagnosis and skin lesion diagnosis, demonstrating that our framework outperformed other state-of-the-art FU frameworks, with an expected speed-up of 10-15 times compared with retraining from scratch. The code and the organized datasets can be found at: https://github.com/dzp2095/FCU.
Authors: Lorenzo Lamberti, Vlad Niculescu, Micha{\l} Barcis, Lorenzo Bellone, Enrico Natalizio, Luca Benini, Daniele Palossi
Abstract: Pocket-sized autonomous nano-drones can revolutionize many robotic use cases, such as visual inspection in narrow, constrained spaces, and ensure safer human-robot interaction due to their tiny form factor and weight -- i.e., tens of grams. This compelling vision is challenged by the high level of intelligence needed aboard, which clashes against the limited computational and storage resources available on PULP (parallel-ultra-low-power) MCU class navigation and mission controllers that can be hosted aboard. This work moves from PULP-Dronet, a State-of-the-Art convolutional neural network for autonomous navigation on nano-drones. We introduce Tiny-PULP-Dronet: a novel methodology to squeeze by more than one order of magnitude model size (50x fewer parameters), and number of operations (27x less multiply-and-accumulate) required to run inference with similar flight performance as PULP-Dronet. This massive reduction paves the way towards affordable multi-tasking on nano-drones, a fundamental requirement for achieving high-level intelligence.
Authors: Gabriele Lozupone, Alessandro Bria, Francesco Fontanella, Claudio De Stefano
Abstract: This study presents an innovative method for Alzheimer's disease diagnosis using 3D MRI designed to enhance the explainability of model decisions. Our approach adopts a soft attention mechanism, enabling 2D CNNs to extract volumetric representations. At the same time, the importance of each slice in decision-making is learned, allowing the generation of a voxel-level attention map to produces an explainable MRI. To test our method and ensure the reproducibility of our results, we chose a standardized collection of MRI data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). On this dataset, our method significantly outperforms state-of-the-art methods in (i) distinguishing AD from cognitive normal (CN) with an accuracy of 0.856 and Matthew's correlation coefficient (MCC) of 0.712, representing improvements of 2.4\% and 5.3\% respectively over the second-best, and (ii) in the prognostic task of discerning stable from progressive mild cognitive impairment (MCI) with an accuracy of 0.725 and MCC of 0.443, showing improvements of 10.2\% and 20.5\% respectively over the second-best. We achieved this prognostic result by adopting a double transfer learning strategy, which enhanced sensitivity to morphological changes and facilitated early-stage AD detection. With voxel-level precision, our method identified which specific areas are being paid attention to, identifying these predominant brain regions: the \emph{hippocampus}, the \emph{amygdala}, the \emph{parahippocampal}, and the \emph{inferior lateral ventricles}. All these areas are clinically associated with AD development. Furthermore, our approach consistently found the same AD-related areas across different cross-validation folds, proving its robustness and precision in highlighting areas that align closely with known pathological markers of the disease.
Authors: Yu Zhe, Jun Sakuma
Abstract: The widespread use of personal data for training machine learning models raises significant privacy concerns, as individuals have limited control over how their public data is subsequently utilized. Availability attacks have emerged as a means for data owners to safeguard their data by desning imperceptible perturbations that degrade model performance when incorporated into training datasets. However, existing availability attacks exhibit limitations in practical applicability, particularly when only a portion of the data can be perturbed. To address this challenge, we propose a novel availability attack approach termed Parameter Matching Attack (PMA). PMA is the first availability attack that works when only a portion of data can be perturbed. PMA optimizes perturbations so that when the model is trained on a mixture of clean and perturbed data, the resulting model will approach a model designed to perform poorly. Experimental results across four datasets demonstrate that PMA outperforms existing methods, achieving significant model performance degradation when a part of the training data is perturbed. Our code is available in the supplementary.
Authors: Tom\'e Albuquerque, Anil Y\"uce, Markus D. Herrmann, Alvaro Gomariz
Abstract: Interpreting machine learning model decisions is crucial for high-risk applications like healthcare. In digital pathology, large whole slide images (WSIs) are decomposed into smaller tiles and tile-derived features are processed by attention-based multiple instance learning (ABMIL) models to predict WSI-level labels. These networks generate tile-specific attention weights, which can be visualized as attention maps for interpretability. However, a standardized evaluation framework for these maps is lacking, questioning their reliability and ability to detect spurious correlations that can mislead models. We herein propose a framework to assess the ability of attention networks to attend to relevant features in digital pathology by creating artificial model confounders and using dedicated interpretability metrics. Models are trained and evaluated on data with tile modifications correlated with WSI labels, enabling the analysis of model sensitivity to artificial confounders and the accuracy of attention maps in highlighting them. Confounders are introduced either through synthetic tile modifications or through tile ablations based on their specific image-based features, with the latter being used to assess more clinically relevant scenarios. We also analyze the impact of varying confounder quantities at both the tile and WSI levels. Our results show that ABMIL models perform as desired within our framework. While attention maps generally highlight relevant regions, their robustness is affected by the type and number of confounders. Our versatile framework has the potential to be used in the evaluation of various methods and the exploration of image-based features driving model predictions, which could aid in biomarker discovery.
Authors: Yeong-Jun Cho
Abstract: In recent years, many semantic segmentation methods have been proposed to predict label of pixels in the scene. In general, we measure area prediction errors or boundary prediction errors for comparing methods. However, there is no intuitive evaluation metric that evaluates both aspects. In this work, we propose a new evaluation measure called weighted Intersection over Union (wIoU) for semantic segmentation. First, it builds a weight map generated from a boundary distance map, allowing weighted evaluation for each pixel based on a boundary importance factor. The proposed wIoU can evaluate both contour and region by setting a boundary importance factor. We validated the effectiveness of wIoU on a dataset of 33 scenes and demonstrated its flexibility. Using the proposed metric, we expect more flexible and intuitive evaluation in semantic segmentation field are possible.
Authors: Hulin Li, Jun Li, Hanbing Wei, Zheng Liu, Zhenfei Zhan, Qiliang Ren
Abstract: Real-time object detection is significant for industrial and research fields. On edge devices, a giant model is difficult to achieve the real-time detecting requirement and a lightweight model built from a large number of the depth-wise separable convolutional could not achieve the sufficient accuracy. We introduce a new lightweight convolutional technique, GSConv, to lighten the model but maintain the accuracy. The GSConv accomplishes an excellent trade-off between the accuracy and speed. Furthermore, we provide a design suggestion based on the GSConv, Slim-Neck (SNs), to achieve a higher computational cost-effectiveness of the real-time detectors. The effectiveness of the SNs was robustly demonstrated in over twenty sets comparative experiments. In particular, the real-time detectors of ameliorated by the SNs obtain the state-of-the-art (70.9% AP50 for the SODA10M at a speed of ~ 100FPS on a Tesla T4) compared with the baselines. Code is available at https://github.com/alanli1997/slim-neck-by-gsconv
Authors: Aashish Anantha Ramakrishnan, Sharon X. Huang, Dongwon Lee
Abstract: Advancements in Text-to-Image synthesis over recent years have focused more on improving the quality of generated samples using datasets with descriptive prompts. However, real-world image-caption pairs present in domains such as news data do not use simple and directly descriptive captions. With captions containing information on both the image content and underlying contextual cues, they become abstractive in nature. In this paper, we launch ANNA, an Abstractive News captioNs dAtaset extracted from online news articles in a variety of different contexts. We explore the capabilities of current Text-to-Image synthesis models to generate news domain-specific images using abstractive captions by benchmarking them on ANNA, in both standard training and transfer learning settings. The generated images are judged on the basis of contextual relevance, visual quality, and perceptual similarity to ground-truth image-caption pairs. Through our experiments, we show that techniques such as transfer learning achieve limited success in understanding abstractive captions but still fail to consistently learn the relationships between content and context features. The Dataset is available at https://github.com/aashish2000/ANNA .
Authors: Shaoxu Li, Ye Pan
Abstract: We present Instant Neural Radiance Fields Stylization, a novel approach for multi-view image stylization for the 3D scene. Our approach models a neural radiance field based on neural graphics primitives, which use a hash table-based position encoder for position embedding. We split the position encoder into two parts, the content and style sub-branches, and train the network for normal novel view image synthesis with the content and style targets. In the inference stage, we execute AdaIN to the output features of the position encoder, with content and style voxel grid features as reference. With the adjusted features, the stylization of novel view images could be obtained. Our method extends the style target from style images to image sets of scenes and does not require additional network training for stylization. Given a set of images of 3D scenes and a style target(a style image or another set of 3D scenes), our method can generate stylized novel views with a consistent appearance at various view angles in less than 10 minutes on modern GPU hardware. Extensive experimental results demonstrate the validity and superiority of our method.
Authors: Lie Ju, Yicheng Wu, Wei Feng, Zhen Yu, Lin Wang, Zhuoting Zhu, Zongyuan Ge
Abstract: Semi-supervised learning (SSL) has attracted much attention since it reduces the expensive costs of collecting adequate well-labeled training data, especially for deep learning methods. However, traditional SSL is built upon an assumption that labeled and unlabeled data should be from the same distribution \textit{e.g.,} classes and domains. However, in practical scenarios, unlabeled data would be from unseen classes or unseen domains, and it is still challenging to exploit them by existing SSL methods. Therefore, in this paper, we proposed a unified framework to leverage these unseen unlabeled data for open-scenario semi-supervised medical image classification. We first design a novel scoring mechanism, called dual-path outliers estimation, to identify samples from unseen classes. Meanwhile, to extract unseen-domain samples, we then apply an effective variational autoencoder (VAE) pre-training. After that, we conduct domain adaptation to fully exploit the value of the detected unseen-domain samples to boost semi-supervised training. We evaluated our proposed framework on dermatology and ophthalmology tasks. Extensive experiments demonstrate our model can achieve superior classification performance in various medical SSL scenarios. The code implementations are accessible at: https://github.com/PyJulie/USSL4MIC.
Authors: Wenxiao Cai, Ke Jin, Jinyan Hou, Cong Guo, Letian Wu, Wankou Yang
Abstract: Semantic segmentation of drone images is critical for various aerial vision tasks as it provides essential semantic details to understand scenes on the ground. Ensuring high accuracy of semantic segmentation models for drones requires access to diverse, large-scale, and high-resolution datasets, which are often scarce in the field of aerial image processing. While existing datasets typically focus on urban scenes and are relatively small, our Varied Drone Dataset (VDD) addresses these limitations by offering a large-scale, densely labeled collection of 400 high-resolution images spanning 7 classes. This dataset features various scenes in urban, industrial, rural, and natural areas, captured from different camera angles and under diverse lighting conditions. We also make new annotations to UDD and UAVid, integrating them under VDD annotation standards, to create the Integrated Drone Dataset (IDD). We train seven state-of-the-art models on drone datasets as baselines. It's expected that our dataset will generate considerable interest in drone image segmentation and serve as a foundation for other drone vision tasks. Datasets are publicly available at \href{our website}{https://github.com/RussRobin/VDD}.
Authors: Patrick Feeney, Michael C. Hughes
Abstract: The information noise-contrastive estimation (InfoNCE) loss function provides the basis of many self-supervised deep learning methods due to its strong empirical results and theoretic motivation. Previous work suggests a supervised contrastive (SupCon) loss to extend InfoNCE to learn from available class labels. This SupCon loss has been widely-used due to reports of good empirical performance. However, in this work we find that the prior SupCon loss formulation has questionable justification because it can encourage some images from the same class to repel one another in the learned embedding space. This problematic intra-class repulsion gets worse as the number of images sharing one class label increases. We propose the Supervised InfoNCE REvisited (SINCERE) loss as a theoretically-justified supervised extension of InfoNCE that eliminates intra-class repulsion. Experiments show that SINCERE leads to better separation of embeddings from different classes and improves transfer learning classification accuracy. We additionally utilize probabilistic modeling to derive an information-theoretic bound that relates SINCERE loss to the symmeterized KL divergence between data-generating distributions for a target class and all other classes.
Authors: Chuan Fang, Yuan Dong, Kunming Luo, Xiaotao Hu, Rakesh Shrestha, Ping Tan
Abstract: Text-driven 3D indoor scene generation is useful for gaming, the film industry, and AR/VR applications. However, existing methods cannot faithfully capture the room layout, nor do they allow flexible editing of individual objects in the room. To address these problems, we present Ctrl-Room, which can generate convincing 3D rooms with designer-style layouts and high-fidelity textures from just a text prompt. Moreover, Ctrl-Room enables versatile interactive editing operations such as resizing or moving individual furniture items. Our key insight is to separate the modeling of layouts and appearance. Our proposed method consists of two stages: a Layout Generation Stage and an Appearance Generation Stage. The Layout Generation Stage trains a text-conditional diffusion model to learn the layout distribution with our holistic scene code parameterization. Next, the Appearance Generation Stage employs a fine-tuned ControlNet to produce a vivid panoramic image of the room guided by the 3D scene layout and text prompt. We thus achieve a high-quality 3D room generation with convincing layouts and lively textures. Benefiting from the scene code parameterization, we can easily edit the generated room model through our mask-guided editing module, without expensive edit-specific training. Extensive experiments on the Structured3D dataset demonstrate that our method outperforms existing methods in producing more reasonable, view-consistent, and editable 3D rooms from natural language prompts.
Authors: Zhengdi Yu, Shaoli Huang, Yongkang Cheng, Tolga Birdal
Abstract: We present SignAvatars, the first large-scale, multi-prompt 3D sign language (SL) motion dataset designed to bridge the communication gap for Deaf and hard-of-hearing individuals. While there has been an exponentially growing number of research regarding digital communication, the majority of existing communication technologies primarily cater to spoken or written languages, instead of SL, the essential communication method for Deaf and hard-of-hearing communities. Existing SL datasets, dictionaries, and sign language production (SLP) methods are typically limited to 2D as annotating 3D models and avatars for SL is usually an entirely manual and labor-intensive process conducted by SL experts, often resulting in unnatural avatars. In response to these challenges, we compile and curate the SignAvatars dataset, which comprises 70,000 videos from 153 signers, totaling 8.34 million frames, covering both isolated signs and continuous, co-articulated signs, with multiple prompts including HamNoSys, spoken language, and words. To yield 3D holistic annotations, including meshes and biomechanically-valid poses of body, hands, and face, as well as 2D and 3D keypoints, we introduce an automated annotation pipeline operating on our large corpus of SL videos. SignAvatars facilitates various tasks such as 3D sign language recognition (SLR) and the novel 3D SL production (SLP) from diverse inputs like text scripts, individual words, and HamNoSys notation. Hence, to evaluate the potential of SignAvatars, we further propose a unified benchmark of 3D SL holistic motion production. We believe that this work is a significant step forward towards bringing the digital world to the Deaf and hard-of-hearing communities as well as people interacting with them.
Authors: Carolina A. M. Heming, Mohamed Abdalla, Shahram Mohanna, Monish Ahluwalia, Linglin Zhang, Hari Trivedi, MinJae Woo, Benjamin Fine, Judy Wawira Gichoya, Leo Anthony Celi, Laleh Seyyed-Kalantari
Abstract: Clinical AI model reporting cards should be expanded to incorporate a broad bias reporting of both social and non-social factors. Non-social factors consider the role of other factors, such as disease dependent, anatomic, or instrument factors on AI model bias, which are essential to ensure safe deployment.
Authors: Kwanyoung Kim, Yujin Oh, Sangjoon Park, Hwa Kyung Byun, Joongyo Lee, Jin Sung Kim, Yong Bae Kim, Jong Chul Ye
Abstract: Recent advances in AI foundation models have significant potential for lightening the clinical workload by mimicking the comprehensive and multi-faceted approaches used by medical professionals. In the field of radiation oncology, the integration of multiple modalities holds great importance, so the opportunity of foundational model is abundant. Inspired by this, here we present RO-LMM, a multi-purpose, comprehensive large multimodal model (LMM) tailored for the field of radiation oncology. This model effectively manages a series of tasks within the clinical workflow, including clinical context summarization, radiation treatment plan suggestion, and plan-guided target volume segmentation by leveraging the capabilities of LMM. In particular, to perform consecutive clinical tasks without error accumulation, we present a novel Consistency Embedding Fine-Tuning (CEFTune) technique, which boosts LMM's robustness to noisy inputs while preserving the consistency of handling clean inputs. We further extend this concept to LMM-driven segmentation framework, leading to a novel Consistency Embedding Segmentation~(CESEG) techniques. Experimental results including multi-centre validation confirm that our RO-LMM with CEFTune and CESEG results in promising performance for multiple clinical tasks with generalization capabilities.
Authors: Raviteja Vemulapalli, Hadi Pouransari, Fartash Faghri, Sachin Mehta, Mehrdad Farajtabar, Mohammad Rastegari, Oncel Tuzel
Abstract: Vision Foundation Models (VFMs) pretrained on massive datasets exhibit impressive performance on various downstream tasks, especially with limited labeled target data. However, due to their high inference compute cost, these models cannot be deployed for many real-world applications. Motivated by this, we ask the following important question, "How can we leverage the knowledge from a large VFM to train a small task-specific model for a new target task with limited labeled training data?", and propose a simple task-oriented knowledge transfer approach as a highly effective solution to this problem. Our experimental results on five target tasks show that the proposed approach outperforms task-agnostic VFM distillation, web-scale CLIP pretraining, supervised ImageNet pretraining, and self-supervised DINO pretraining by up to 11.6%, 22.1%, 13.7%, and 29.8%, respectively. Furthermore, the proposed approach also demonstrates up to 9x, 4x and 15x reduction in pretraining compute cost when compared to task-agnostic VFM distillation, ImageNet pretraining and DINO pretraining, respectively, while outperforming them. We also show that the dataset used for transferring knowledge has a significant effect on the final target task performance, and introduce a retrieval-augmented knowledge transfer strategy that uses web-scale image retrieval to curate effective transfer sets.
Authors: Zhiyu Pan, Yongjie Duan, Jianjiang Feng, Jie Zhou
Abstract: In fingerprint matching, fixed-length descriptors generally offer greater efficiency compared to minutiae set, but the recognition accuracy is not as good as that of the latter. Although much progress has been made in deep learning based fixed-length descriptors recently, they often fall short when dealing with incomplete or partial fingerprints, diverse fingerprint poses, and significant background noise. In this paper, we propose a three-dimensional representation called Fixed-length Dense Descriptor (FDD) for efficient fingerprint matching. FDD features great spatial properties, enabling it to capture the spatial relationships of the original fingerprints, thereby enhancing interpretability and robustness. Our experiments on various fingerprint datasets reveal that FDD outperforms other fixed-length descriptors, especially in matching fingerprints of different areas, cross-modal fingerprint matching, and fingerprint matching with background noise.
Authors: Daniel Duckworth, Peter Hedman, Christian Reiser, Peter Zhizhin, Jean-Fran\c{c}ois Thibert, Mario Lu\v{c}i\'c, Richard Szeliski, Jonathan T. Barron
Abstract: Recent techniques for real-time view synthesis have rapidly advanced in fidelity and speed, and modern methods are capable of rendering near-photorealistic scenes at interactive frame rates. At the same time, a tension has arisen between explicit scene representations amenable to rasterization and neural fields built on ray marching, with state-of-the-art instances of the latter surpassing the former in quality while being prohibitively expensive for real-time applications. In this work, we introduce SMERF, a view synthesis approach that achieves state-of-the-art accuracy among real-time methods on large scenes with footprints up to 300 m$^2$ at a volumetric resolution of 3.5 mm$^3$. Our method is built upon two primary contributions: a hierarchical model partitioning scheme, which increases model capacity while constraining compute and memory consumption, and a distillation training strategy that simultaneously yields high fidelity and internal consistency. Our approach enables full six degrees of freedom (6DOF) navigation within a web browser and renders in real-time on commodity smartphones and laptops. Extensive experiments show that our method exceeds the current state-of-the-art in real-time novel view synthesis by 0.78 dB on standard benchmarks and 1.78 dB on large scenes, renders frames three orders of magnitude faster than state-of-the-art radiance field models, and achieves real-time performance across a wide variety of commodity devices, including smartphones. We encourage readers to explore these models interactively at our project website: https://smerf-3d.github.io.
Authors: Dingkun Yan, Liang Yuan, Erwin Wu, Yuma Nishioka, Issei Fujishiro, Suguru Saito
Abstract: Diffusion models have recently demonstrated their effectiveness in generating extremely high-quality images and are now utilized in a wide range of applications, including automatic sketch colorization. Although many methods have been developed for guided sketch colorization, there has been limited exploration of the potential conflicts between image prompts and sketch inputs, which can lead to severe deterioration in the results. Therefore, this paper exhaustively investigates reference-based sketch colorization models that aim to colorize sketch images using reference color images. We specifically investigate two critical aspects of reference-based diffusion models: the "distribution problem", which is a major shortcoming compared to text-based counterparts, and the capability in zero-shot sequential text-based manipulation. We introduce two variations of an image-guided latent diffusion model utilizing different image tokens from the pre-trained CLIP image encoder and propose corresponding manipulation methods to adjust their results sequentially using weighted text inputs. We conduct comprehensive evaluations of our models through qualitative and quantitative experiments as well as a user study.
Authors: Chunlei Peng, Boyu Wang, Decheng Liu, Nannan Wang, Ruimin Hu, Xinbo Gao
Abstract: Cloth-changing person re-identification (CC-ReID) aims to match persons who change clothes over long periods. The key challenge in CC-ReID is to extract clothing-independent features, such as face, hairstyle, body shape, and gait. Current research mainly focuses on modeling body shape using multi-modal biological features (such as silhouettes and sketches). However, it does not fully leverage the personal description information hidden in the original RGB image. Considering that there are certain attribute descriptions which remain unchanged after the changing of cloth, we propose a Masked Attribute Description Embedding (MADE) method that unifies personal visual appearance and attribute description for CC-ReID. Specifically, handling variable clothing-sensitive information, such as color and type, is challenging for effective modeling. To address this, we mask the clothing and color information in the personal attribute description extracted through an attribute detection model. The masked attribute description is then connected and embedded into Transformer blocks at various levels, fusing it with the low-level to high-level features of the image. This approach compels the model to discard clothing information. Experiments are conducted on several CC-ReID benchmarks, including PRCC, LTCC, Celeb-reID-light, and LaST. Results demonstrate that MADE effectively utilizes attribute description, enhancing cloth-changing person re-identification performance, and compares favorably with state-of-the-art methods. The code is available at https://github.com/moon-wh/MADE.
Authors: Hao Wang, Shuhei Kurita, Shuichiro Shimizu, Daisuke Kawahara
Abstract: Audio-visual speech recognition (AVSR) is a multimodal extension of automatic speech recognition (ASR), using video as a complement to audio. In AVSR, considerable efforts have been directed at datasets for facial features such as lip-readings, while they often fall short in evaluating the image comprehension capabilities in broader contexts. In this paper, we construct SlideAVSR, an AVSR dataset using scientific paper explanation videos. SlideAVSR provides a new benchmark where models transcribe speech utterances with texts on the slides on the presentation recordings. As technical terminologies that are frequent in paper explanations are notoriously challenging to transcribe without reference texts, our SlideAVSR dataset spotlights a new aspect of AVSR problems. As a simple yet effective baseline, we propose DocWhisper, an AVSR model that can refer to textual information from slides, and confirm its effectiveness on SlideAVSR.
Authors: Yonghao Yu, Shunan Zhu, Huai Qin, Haorui Li, Jinglu Hu
Abstract: Witnessing the evolution of text-to-image diffusion models, significant strides have been made in text-to-3D generation. Currently, two primary paradigms dominate the field of text-to-3D: the feed-forward generation solutions, capable of swiftly producing 3D assets but often yielding coarse results, and the Score Distillation Sampling (SDS) based solutions, known for generating high-fidelity 3D assets albeit at a slower pace. The synergistic integration of these methods holds substantial promise for advancing 3D generation techniques. In this paper, we present BoostDream, a highly efficient plug-and-play 3D refining method designed to transform coarse 3D assets into high-quality. The BoostDream framework comprises three distinct processes: (1) We introduce 3D model distillation that fits differentiable representations from the 3D assets obtained through feed-forward generation. (2) A novel multi-view SDS loss is designed, which utilizes a multi-view aware 2D diffusion model to refine the 3D assets. (3) We propose to use prompt and multi-view consistent normal maps as guidance in refinement.Our extensive experiment is conducted on different differentiable 3D representations, revealing that BoostDream excels in generating high-quality 3D assets rapidly, overcoming the Janus problem compared to conventional SDS-based methods. This breakthrough signifies a substantial advancement in both the efficiency and quality of 3D generation processes.
Authors: Lifan Liang
Abstract: In this paper, we showed that the feature map of a convolution layer is equivalent to the unnormalized log posterior of a special kind of Gaussian mixture for image modeling. Then we expanded the model to drive diverse features and proposed a corresponding EM algorithm to learn the model. Learning convolution weights using this approach is efficient, guaranteed to converge, and does not need supervised information. Code is available at: https://github.com/LifanLiang/CALM.
Authors: Yuanxing Duan, Fangyin Wei, Qiyu Dai, Yuhang He, Wenzheng Chen, Baoquan Chen
Abstract: We consider the problem of novel-view synthesis (NVS) for dynamic scenes. Recent neural approaches have accomplished exceptional NVS results for static 3D scenes, but extensions to 4D time-varying scenes remain non-trivial. Prior efforts often encode dynamics by learning a canonical space plus implicit or explicit deformation fields, which struggle in challenging scenarios like sudden movements or generating high-fidelity renderings. In this paper, we introduce 4D Gaussian Splatting (4DRotorGS), a novel method that represents dynamic scenes with anisotropic 4D XYZT Gaussians, inspired by the success of 3D Gaussian Splatting in static scenes. We model dynamics at each timestamp by temporally slicing the 4D Gaussians, which naturally compose dynamic 3D Gaussians and can be seamlessly projected into images. As an explicit spatial-temporal representation, 4DRotorGS demonstrates powerful capabilities for modeling complicated dynamics and fine details--especially for scenes with abrupt motions. We further implement our temporal slicing and splatting techniques in a highly optimized CUDA acceleration framework, achieving real-time inference rendering speeds of up to 277 FPS on an RTX 3090 GPU and 583 FPS on an RTX 4090 GPU. Rigorous evaluations on scenes with diverse motions showcase the superior efficiency and effectiveness of 4DRotorGS, which consistently outperforms existing methods both quantitatively and qualitatively.
Authors: Shuchen Xue, Zhaoqiang Liu, Fei Chen, Shifeng Zhang, Tianyang Hu, Enze Xie, Zhenguo Li
Abstract: Diffusion probabilistic models (DPMs) have shown remarkable performance in high-resolution image synthesis, but their sampling efficiency is still to be desired due to the typically large number of sampling steps. Recent advancements in high-order numerical ODE solvers for DPMs have enabled the generation of high-quality images with much fewer sampling steps. While this is a significant development, most sampling methods still employ uniform time steps, which is not optimal when using a small number of steps. To address this issue, we propose a general framework for designing an optimization problem that seeks more appropriate time steps for a specific numerical ODE solver for DPMs. This optimization problem aims to minimize the distance between the ground-truth solution to the ODE and an approximate solution corresponding to the numerical solver. It can be efficiently solved using the constrained trust region method, taking less than $15$ seconds. Our extensive experiments on both unconditional and conditional sampling using pixel- and latent-space DPMs demonstrate that, when combined with the state-of-the-art sampling method UniPC, our optimized time steps significantly improve image generation performance in terms of FID scores for datasets such as CIFAR-10 and ImageNet, compared to using uniform time steps.
Authors: Yijin Huang, Pujin Cheng, Roger Tam, Xiaoying Tang
Abstract: Parameter-efficient transfer learning (PETL) is proposed as a cost-effective way to transfer pre-trained models to downstream tasks, avoiding the high cost of updating entire large-scale pre-trained models (LPMs). In this work, we present Fine-grained Prompt Tuning (FPT), a novel PETL method for medical image classification. FPT significantly reduces memory consumption compared to other PETL methods, especially in high-resolution input contexts. To achieve this, we first freeze the weights of the LPM and construct a learnable lightweight side network. The frozen LPM takes high-resolution images as input to extract fine-grained features, while the side network is fed low-resolution images to reduce memory usage. To allow the side network to access pre-trained knowledge, we introduce fine-grained prompts that summarize information from the LPM through a fusion module. Important tokens selection and preloading techniques are employed to further reduce training cost and memory requirements. We evaluate FPT on four medical datasets with varying sizes, modalities, and complexities. Experimental results demonstrate that FPT achieves comparable performance to fine-tuning the entire LPM while using only 1.8% of the learnable parameters and 13% of the memory costs of an encoder ViT-B model with a 512 x 512 input resolution.
Authors: Zijian Wu, Adam Schmidt, Peter Kazanzides, Septimiu E. Salcudean
Abstract: The Segment Anything Model (SAM) is a powerful vision foundation model that is revolutionizing the traditional paradigm of segmentation. Despite this, a reliance on prompting each frame and large computational cost limit its usage in robotically assisted surgery. Applications, such as augmented reality guidance, require little user intervention along with efficient inference to be usable clinically. In this study, we address these limitations by adopting lightweight SAM variants to meet the efficiency requirement and employing fine-tuning techniques to enhance their generalization in surgical scenes. Recent advancements in Tracking Any Point (TAP) have shown promising results in both accuracy and efficiency, particularly when points are occluded or leave the field of view. Inspired by this progress, we present a novel framework that combines an online point tracker with a lightweight SAM model that is fine-tuned for surgical instrument segmentation. Sparse points within the region of interest are tracked and used to prompt SAM throughout the video sequence, providing temporal consistency. The quantitative results surpass the state-of-the-art semi-supervised video object segmentation method XMem on the EndoVis 2015 dataset with 84.8 IoU and 91.0 Dice. Our method achieves promising performance that is comparable to XMem and transformer-based fully supervised segmentation methods on ex vivo UCL dVRK and in vivo CholecSeg8k datasets. In addition, the proposed method shows promising zero-shot generalization ability on the label-free STIR dataset. In terms of efficiency, we tested our method on a single GeForce RTX 4060/4090 GPU respectively, achieving an over 25/90 FPS inference speed. Code is available at: https://github.com/wuzijian1997/SIS-PT-SAM
Authors: Yizhe Xiong, Hui Chen, Tianxiang Hao, Zijia Lin, Jungong Han, Yuesong Zhang, Guoxin Wang, Yongjun Bao, Guiguang Ding
Abstract: Recently, the scale of transformers has grown rapidly, which introduces considerable challenges in terms of training overhead and inference efficiency in the scope of task adaptation. Existing works, namely Parameter-Efficient Fine-Tuning (PEFT) and model compression, have separately investigated the challenges. However, PEFT cannot guarantee the inference efficiency of the original backbone, especially for large-scale models. Model compression requires significant training costs for structure searching and re-training. Consequently, a simple combination of them cannot guarantee accomplishing both training efficiency and inference efficiency with minimal costs. In this paper, we propose a novel Parallel Yielding Re-Activation (PYRA) method for such a challenge of training-inference efficient task adaptation. PYRA first utilizes parallel yielding adaptive weights to comprehensively perceive the data distribution in downstream tasks. A re-activation strategy for token modulation is then applied for tokens to be merged, leading to calibrated token features. Extensive experiments demonstrate that PYRA outperforms all competing methods under both low compression rate and high compression rate, demonstrating its effectiveness and superiority in maintaining both training efficiency and inference efficiency for large-scale foundation models. Our code will be released to the public.
Authors: Ziying Song, Lei Yang, Shaoqing Xu, Lin Liu, Dongyang Xu, Caiyan Jia, Feiyang Jia, Li Wang
Abstract: Integrating LiDAR and camera information into Bird's-Eye-View (BEV) representation has emerged as a crucial aspect of 3D object detection in autonomous driving. However, existing methods are susceptible to the inaccurate calibration relationship between LiDAR and the camera sensor. Such inaccuracies result in errors in depth estimation for the camera branch, ultimately causing misalignment between LiDAR and camera BEV features. In this work, we propose a robust fusion framework called Graph BEV. Addressing errors caused by inaccurate point cloud projection, we introduce a Local Align module that employs neighbor-aware depth features via Graph matching. Additionally, we propose a Global Align module to rectify the misalignment between LiDAR and camera BEV features. Our Graph BEV framework achieves state-of-the-art performance, with an mAP of 70.1\%, surpassing BEV Fusion by 1.6\% on the nuscenes validation set. Importantly, our Graph BEV outperforms BEV Fusion by 8.3\% under conditions with misalignment noise.
Authors: Jingyu Lin, Jiaqi Gu, Bojian Wu, Lubin Fan, Renjie Chen, Ligang Liu, Jieping Ye
Abstract: We introduce a novel neural volumetric pose feature, termed PoseMap, designed to enhance camera localization by encapsulating the information between images and the associated camera poses. Our framework leverages an Absolute Pose Regression (APR) architecture, together with an augmented NeRF module. This integration not only facilitates the generation of novel views to enrich the training dataset but also enables the learning of effective pose features. Additionally, we extend our architecture for self-supervised online alignment, allowing our method to be used and fine-tuned for unlabelled images within a unified framework. Experiments demonstrate that our method achieves 14.28% and 20.51% performance gain on average in indoor and outdoor benchmark scenes, outperforming existing APR methods with state-of-the-art accuracy.
Authors: Zeliang Zhang, Mingqian Feng, Jinyang Jiang, Rongyi Zhu, Yijie Peng, Chenliang Xu
Abstract: Gradient-based saliency maps are widely used to explain deep neural network decisions. However, as models become deeper and more black-box, such as in closed-source APIs like ChatGPT, computing gradients become challenging, hindering conventional explanation methods. In this work, we introduce a novel unified framework for estimating gradients in black-box settings and generating saliency maps to interpret model decisions. We employ the likelihood ratio method to estimate output-to-input gradients and utilize them for saliency map generation. Additionally, we propose blockwise computation techniques to enhance estimation accuracy. Extensive experiments in black-box settings validate the effectiveness of our method, demonstrating accurate gradient estimation and explainability of generated saliency maps. Furthermore, we showcase the scalability of our approach by applying it to explain GPT-Vision, revealing the continued relevance of gradient-based explanation methods in the era of large, closed-source, and black-box models.
Authors: Jiuming Liu, Dong Zhuo, Zhiheng Feng, Siting Zhu, Chensheng Peng, Zhe Liu, Hesheng Wang
Abstract: Information inside visual and LiDAR data is well complementary derived from the fine-grained texture of images and massive geometric information in point clouds. However, it remains challenging to explore effective visual-LiDAR fusion, mainly due to the intrinsic data structure inconsistency between two modalities: Images are regular and dense, but LiDAR points are unordered and sparse. To address the problem, we propose a local-to-global fusion network with bi-directional structure alignment. To obtain locally fused features, we project points onto image plane as cluster centers and cluster image pixels around each center. Image pixels are pre-organized as pseudo points for image-to-point structure alignment. Then, we convert points to pseudo images by cylindrical projection (point-to-image structure alignment) and perform adaptive global feature fusion between point features with local fused features. Our method achieves state-of-the-art performance on KITTI odometry and FlyingThings3D scene flow datasets compared to both single-modal and multi-modal methods. Codes will be released at https://github.com/IRMVLab/DVLO.
Authors: Shaoxu Li
Abstract: Image style transfer aims to imbue digital imagery with the distinctive attributes of style targets, such as colors, brushstrokes, shapes, whilst concurrently preserving the semantic integrity of the content. Despite the advancements in arbitrary style transfer methods, a prevalent challenge remains the delicate equilibrium between content semantics and style attributes. Recent developments in large-scale text-to-image diffusion models have heralded unprecedented synthesis capabilities, albeit at the expense of relying on extensive and often imprecise textual descriptions to delineate artistic styles. Addressing these limitations, this paper introduces DiffStyler, a novel approach that facilitates efficient and precise arbitrary image style transfer. DiffStyler lies the utilization of a text-to-image Stable Diffusion model-based LoRA to encapsulate the essence of style targets. This approach, coupled with strategic cross-LoRA feature and attention injection, guides the style transfer process. The foundation of our methodology is rooted in the observation that LoRA maintains the spatial feature consistency of UNet, a discovery that further inspired the development of a mask-wise style transfer technique. This technique employs masks extracted through a pre-trained FastSAM model, utilizing mask prompts to facilitate feature fusion during the denoising process, thereby enabling localized style transfer that preserves the original image's unaffected regions. Moreover, our approach accommodates multiple style targets through the use of corresponding masks. Through extensive experimentation, we demonstrate that DiffStyler surpasses previous methods in achieving a more harmonious balance between content preservation and style integration.
Authors: Tao Hu, Fangzhou Hong, Ziwei Liu
Abstract: Recent 3D human generative models have achieved remarkable progress by learning 3D-aware GANs from 2D images. However, existing 3D human generative methods model humans in a compact 1D latent space, ignoring the articulated structure and semantics of human body topology. In this paper, we explore more expressive and higher-dimensional latent space for 3D human modeling and propose StructLDM, a diffusion-based unconditional 3D human generative model, which is learned from 2D images. StructLDM solves the challenges imposed due to the high-dimensional growth of latent space with three key designs: 1) A semantic structured latent space defined on the dense surface manifold of a statistical human body template. 2) A structured 3D-aware auto-decoder that factorizes the global latent space into several semantic body parts parameterized by a set of conditional structured local NeRFs anchored to the body template, which embeds the properties learned from the 2D training data and can be decoded to render view-consistent humans under different poses and clothing styles. 3) A structured latent diffusion model for generative human appearance sampling. Extensive experiments validate StructLDM's state-of-the-art generation performance and illustrate the expressiveness of the structured latent space over the well-adopted 1D latent space. Notably, StructLDM enables different levels of controllable 3D human generation and editing, including pose/view/shape control, and high-level tasks including compositional generations, part-aware clothing editing, 3D virtual try-on, etc. Our project page is at: https://taohuumd.github.io/projects/StructLDM/.
Authors: Zhida Zhang, Jie Cao, Wenkui Yang, Qihang Fan, Kai Zhou, Ran He
Abstract: The transformer networks are extensively utilized in face forgery detection due to their scalability across large datasets.Despite their success, transformers face challenges in balancing the capture of global context, which is crucial for unveiling forgery clues, with computational complexity.To mitigate this issue, we introduce Band-Attention modulated RetNet (BAR-Net), a lightweight network designed to efficiently process extensive visual contexts while avoiding catastrophic forgetting.Our approach empowers the target token to perceive global information by assigning differential attention levels to tokens at varying distances. We implement self-attention along both spatial axes, thereby maintaining spatial priors and easing the computational burden.Moreover, we present the adaptive frequency Band-Attention Modulation mechanism, which treats the entire Discrete Cosine Transform spectrogram as a series of frequency bands with learnable weights.Together, BAR-Net achieves favorable performance on several face forgery datasets, outperforming current state-of-the-art methods.
Authors: Zander W. Blasingame, Chen Liu
Abstract: Morphing attacks are an emerging threat to state-of-the-art Face Recognition (FR) systems, which aim to create a single image that contains the biometric information of multiple identities. Diffusion Morphs (DiM) are a recently proposed morphing attack that has achieved state-of-the-art performance for representation-based morphing attacks. However, none of the existing research on DiMs have leveraged the iterative nature of DiMs and left the DiM model as a black box, treating it no differently than one would a Generative Adversarial Network (GAN) or Varational AutoEncoder (VAE). We propose a greedy strategy on the iterative sampling process of DiM models which searches for an optimal step guided by an identity-based heuristic function. We compare our proposed algorithm against ten other state-of-the-art morphing algorithms using the open-source SYN-MAD 2022 competition dataset. We find that our proposed algorithm is unreasonably effective, fooling all of the tested FR systems with an MMPMR of 100%, outperforming all other morphing algorithms compared.
Authors: Rui Zhao, Bin Shi, Jianfei Ruan, Tianze Pan, Bo Dong
Abstract: In noisy label learning, estimating noisy class posteriors plays a fundamental role for developing consistent classifiers, as it forms the basis for estimating clean class posteriors and the transition matrix. Existing methods typically learn noisy class posteriors by training a classification model with noisy labels. However, when labels are incorrect, these models may be misled to overemphasize the feature parts that do not reflect the instance characteristics, resulting in significant errors in estimating noisy class posteriors. To address this issue, this paper proposes to augment the supervised information with part-level labels, encouraging the model to focus on and integrate richer information from various parts. Specifically, our method first partitions features into distinct parts by cropping instances, yielding part-level labels associated with these various parts. Subsequently, we introduce a novel single-to-multiple transition matrix to model the relationship between the noisy and part-level labels, which incorporates part-level labels into a classifier-consistent framework. Utilizing this framework with part-level labels, we can learn the noisy class posteriors more precisely by guiding the model to integrate information from various parts, ultimately improving the classification performance. Our method is theoretically sound, while experiments show that it is empirically effective in synthetic and real-world noisy benchmarks.
Authors: Ji Ma, Wei Suo, Peng Wang, Yanning Zhang
Abstract: Vision-Language Instruction Tuning (VLIT) is a critical training phase for Large Vision-Language Models (LVLMs). With the improving capabilities of open-source LVLMs, researchers have increasingly turned to generate VLIT data by using open-source LVLMs and achieved significant progress. However, such data generation approaches are bottlenecked by the following challenges: 1) Since multi-modal models tend to be influenced by prior language knowledge, directly using LVLMs to generate VLIT data would inevitably lead to low content relevance between generated data and images. 2) To improve the ability of the models to generate VLIT data, previous methods have incorporated an additional training phase to boost the generative capacity. This process hurts the generalization of the models to unseen inputs (i.e., "exposure bias" problem). In this paper, we propose a new Content Correlated VLIT data generation via Contrastive Learning (C3L). Specifically, we design a new content relevance module which enhances the content relevance between VLIT data and images by computing Image Instruction Correspondence Scores S(I2C). Moreover, a contrastive learning module is introduced to further boost the VLIT data generation capability of the LVLMs. A large number of automatic measures on four benchmarks show the effectiveness of our method.
Authors: Andres Hernandez, Zhongqi Miao, Luisa Vargas, Rahul Dodhia, Pablo Arbelaez, Juan M. Lavista Ferres
Abstract: The alarming decline in global biodiversity, driven by various factors, underscores the urgent need for large-scale wildlife monitoring. In response, scientists have turned to automated deep learning methods for data processing in wildlife monitoring. However, applying these advanced methods in real-world scenarios is challenging due to their complexity and the need for specialized knowledge, primarily because of technical challenges and interdisciplinary barriers. To address these challenges, we introduce Pytorch-Wildlife, an open-source deep learning platform built on PyTorch. It is designed for creating, modifying, and sharing powerful AI models. This platform emphasizes usability and accessibility, making it accessible to individuals with limited or no technical background. It also offers a modular codebase to simplify feature expansion and further development. Pytorch-Wildlife offers an intuitive, user-friendly interface, accessible through local installation or Hugging Face, for animal detection and classification in images and videos. As two real-world applications, Pytorch-Wildlife has been utilized to train animal classification models for species recognition in the Amazon Rainforest and for invasive opossum recognition in the Galapagos Islands. The Opossum model achieves 98% accuracy, and the Amazon model has 92% recognition accuracy for 36 animals in 90% of the data. As Pytorch-Wildlife evolves, we aim to integrate more conservation tasks, addressing various environmental challenges. Pytorch-Wildlife is available at https://github.com/microsoft/CameraTraps.
Authors: Yuan Shen, Duygu Ceylan, Paul Guerrero, Zexiang Xu, Niloy J. Mitra, Shenlong Wang, Anna Fr\"uhst\"uck
Abstract: We present a simple, modular, and generic method that upsamples coarse 3D models by adding geometric and appearance details. While generative 3D models now exist, they do not yet match the quality of their counterparts in image and video domains. We demonstrate that it is possible to directly repurpose existing (pretrained) video models for 3D super-resolution and thus sidestep the problem of the shortage of large repositories of high-quality 3D training models. We describe how to repurpose video upsampling models, which are not 3D consistent, and combine them with 3D consolidation to produce 3D-consistent results. As output, we produce high quality Gaussian Splat models, which are object centric and effective. Our method is category agnostic and can be easily incorporated into existing 3D workflows. We evaluate our proposed SuperGaussian on a variety of 3D inputs, which are diverse both in terms of complexity and representation (e.g., Gaussian Splats or NeRFs), and demonstrate that our simple method significantly improves the fidelity of the final 3D models. Check our project website for details: supergaussian.github.io
Authors: Benedikt Alkin, Maximilian Beck, Korbinian P\"oppel, Sepp Hochreiter, Johannes Brandstetter
Abstract: Transformers are widely used as generic backbones in computer vision, despite initially introduced for natural language processing. Recently, the Long Short-Term Memory (LSTM) has been extended to a scalable and performant architecture - the xLSTM - which overcomes long-standing LSTM limitations via exponential gating and parallelizable matrix memory structure. In this report, we introduce Vision-LSTM (ViL), an adaption of the xLSTM building blocks to computer vision. ViL comprises a stack of xLSTM blocks where odd blocks process the sequence of patch tokens from top to bottom while even blocks go from bottom to top. Experiments show that ViL holds promise to be further deployed as new generic backbone for computer vision architectures.
Authors: Hang Yao, Ming Liu, Haolin Wang, Zhicun Yin, Zifei Yan, Xiaopeng Hong, Wangmeng Zuo
Abstract: Diffusion models have shown superior performance on unsupervised anomaly detection tasks. Since trained with normal data only, diffusion models tend to reconstruct normal counterparts of test images with certain noises added. However, these methods treat all potential anomalies equally, which may cause two main problems. From the global perspective, the difficulty of reconstructing images with different anomalies is uneven. Therefore, instead of utilizing the same setting for all samples, we propose to predict a particular denoising step for each sample by evaluating the difference between image contents and the priors extracted from diffusion models. From the local perspective, reconstructing abnormal regions differs from normal areas even in the same image. Theoretically, the diffusion model predicts a noise for each step, typically following a standard Gaussian distribution. However, due to the difference between the anomaly and its potential normal counterpart, the predicted noise in abnormal regions will inevitably deviate from the standard Gaussian distribution. To this end, we propose introducing synthetic abnormal samples in training to encourage the diffusion models to break through the limitation of standard Gaussian distribution, and a spatial-adaptive feature fusion scheme is utilized during inference. With the above modifications, we propose a global and local adaptive diffusion model (abbreviated to GLAD) for unsupervised anomaly detection, which introduces appealing flexibility and achieves anomaly-free reconstruction while retaining as much normal information as possible. Extensive experiments are conducted on three commonly used anomaly detection datasets (MVTec-AD, MPDD, and VisA) and a printed circuit board dataset (PCB-Bank) we integrated, showing the effectiveness of the proposed method.
Authors: Fei Wang, Xingyu Fu, James Y. Huang, Zekun Li, Qin Liu, Xiaogeng Liu, Mingyu Derek Ma, Nan Xu, Wenxuan Zhou, Kai Zhang, Tianyi Lorena Yan, Wenjie Jacky Mo, Hsiang-Hui Liu, Pan Lu, Chunyuan Li, Chaowei Xiao, Kai-Wei Chang, Dan Roth, Sheng Zhang, Hoifung Poon, Muhao Chen
Abstract: We introduce MuirBench, a comprehensive benchmark that focuses on robust multi-image understanding capabilities of multimodal LLMs. MuirBench consists of 12 diverse multi-image tasks (e.g., scene understanding, ordering) that involve 10 categories of multi-image relations (e.g., multiview, temporal relations). Comprising 11,264 images and 2,600 multiple-choice questions, MuirBench is created in a pairwise manner, where each standard instance is paired with an unanswerable variant that has minimal semantic differences, in order for a reliable assessment. Evaluated upon 20 recent multi-modal LLMs, our results reveal that even the best-performing models like GPT-4o and Gemini Pro find it challenging to solve MuirBench, achieving 68.0% and 49.3% in accuracy. Open-source multimodal LLMs trained on single images can hardly generalize to multi-image questions, hovering below 33.3% in accuracy. These results highlight the importance of MuirBench in encouraging the community to develop multimodal LLMs that can look beyond a single image, suggesting potential pathways for future improvements.
Authors: Weiqing Xiao
Abstract: Both uncertainty-assisted and iteration-based methods have achieved great success in stereo matching. However, existing uncertainty estimation methods take a single image and the corresponding disparity as input, which imposes higher demands on the estimation network. In this paper, we propose Cost volume-based disparity Uncertainty Estimation (UEC). Based on the rich similarity information in the cost volume coming from the image pairs, the proposed UEC can achieve competitive performance with low computational cost. Secondly, we propose two methods of uncertainty-assisted disparity estimation, Uncertainty-based Disparity Rectification (UDR) and Uncertainty-based Disparity update Conditioning (UDC). These two methods optimise the disparity update process of the iterative-based approach without adding extra parameters. In addition, we propose Disparity Rectification loss that significantly improves the accuracy of small amount of disparity updates. We present a high-performance stereo architecture, DR Stereo, which is a combination of the proposed methods. Experimental results from SceneFlow, KITTI, Middlebury 2014, and ETH3D show that DR-Stereo achieves very competitive disparity estimation performance.
Authors: Guangkun Nie, Qinghao Zhao, Gongzheng Tang, Jun Li, Shenda Hong
Abstract: Photoplethysmography (PPG) is emerging as a crucial tool for monitoring human hemodynamics, with recent studies highlighting its potential in assessing vascular aging through deep learning. However, real-world age distributions are often imbalanced, posing significant challenges for deep learning models. In this paper, we introduce a novel, simple, and effective loss function named the Dist Loss to address deep imbalanced regression tasks. We trained a one-dimensional convolutional neural network (Net1D) incorporating the Dist Loss on the extensive UK Biobank dataset (n=502,389) to estimate vascular age from PPG signals and validate its efficacy in characterizing cardiovascular health. The model's performance was validated on a 40% held-out test set, achieving state-of-the-art results, especially in regions with small sample sizes. Furthermore, we divided the population into three subgroups based on the difference between predicted vascular age and chronological age: less than -10 years, between -10 and 10 years, and greater than 10 years. We analyzed the relationship between predicted vascular age and several cardiovascular events over a follow-up period of up to 10 years, including death, coronary heart disease, and heart failure. Our results indicate that the predicted vascular age has significant potential to reflect an individual's cardiovascular health status. Our code will be available at https://github.com/Ngk03/AI-vascular-age.
Authors: Xiaowen Ma, Rongrong Lian, Zhenkai Wu, Hongbo Guo, Mengting Ma, Sensen Wu, Zhenhong Du, Siyang Song, Wei Zhang
Abstract: Remote sensing images usually characterized by complex backgrounds, scale and orientation variations, and large intra-class variance. General semantic segmentation methods usually fail to fully investigate the above issues, and thus their performances on remote sensing image segmentation are limited. In this paper, we propose our LOGCAN++, a semantic segmentation model customized for remote sensing images, which is made up of a Global Class Awareness (GCA) module and several Local Class Awareness (LCA) modules. The GCA module captures global representations for class-level context modeling to reduce the interference of background noise. The LCA module generates local class representations as intermediate perceptual elements to indirectly associate pixels with the global class representations, targeting at dealing with the large intra-class variance problem. In particular, we introduce affine transformations in the LCA module for adaptive extraction of local class representations to effectively tolerate scale and orientation variations in remotely sensed images. Extensive experiments on three benchmark datasets show that our LOGCAN++ outperforms current mainstream general and remote sensing semantic segmentation methods and achieves a better trade-off between speed and accuracy. Code is available at https://github.com/xwmaxwma/rssegmentation.
Authors: Georgios Tziafas, Yucheng Xu, Zhibin Li, Hamidreza Kasaei
Abstract: Grounding natural language to the physical world is a ubiquitous topic with a wide range of applications in computer vision and robotics. Recently, 2D vision-language models such as CLIP have been widely popularized, due to their impressive capabilities for open-vocabulary grounding in 2D images. Recent works aim to elevate 2D CLIP features to 3D via feature distillation, but either learn neural fields that are scene-specific and hence lack generalization, or focus on indoor room scan data that require access to multiple camera views, which is not practical in robot manipulation scenarios. Additionally, related methods typically fuse features at pixel-level and assume that all camera views are equally informative. In this work, we show that this approach leads to sub-optimal 3D features, both in terms of grounding accuracy, as well as segmentation crispness. To alleviate this, we propose a multi-view feature fusion strategy that employs object-centric priors to eliminate uninformative views based on semantic information, and fuse features at object-level via instance segmentation masks. To distill our object-centric 3D features, we generate a large-scale synthetic multi-view dataset of cluttered tabletop scenes, spawning 15k scenes from over 3300 unique object instances, which we make publicly available. We show that our method reconstructs 3D CLIP features with improved grounding capacity and spatial consistency, while doing so from single-view RGB-D, thus departing from the assumption of multiple camera views at test time. Finally, we show that our approach can generalize to novel tabletop domains and be re-purposed for 3D instance segmentation without fine-tuning, and demonstrate its utility for language-guided robotic grasping in clutter
Authors: Siyuan Liang, Jiawei Liang, Tianyu Pang, Chao Du, Aishan Liu, Ee-Chien Chang, Xiaochun Cao
Abstract: Instruction tuning enhances large vision-language models (LVLMs) but raises security risks through potential backdoor attacks due to their openness. Previous backdoor studies focus on enclosed scenarios with consistent training and testing instructions, neglecting the practical domain gaps that could affect attack effectiveness. This paper empirically examines the generalizability of backdoor attacks during the instruction tuning of LVLMs for the first time, revealing certain limitations of most backdoor strategies in practical scenarios. We quantitatively evaluate the generalizability of six typical backdoor attacks on image caption benchmarks across multiple LVLMs, considering both visual and textual domain offsets. Our findings indicate that attack generalizability is positively correlated with the backdoor trigger's irrelevance to specific images/models and the preferential correlation of the trigger pattern. Additionally, we modify existing backdoor attacks based on the above key observations, demonstrating significant improvements in cross-domain scenario generalizability (+86% attack success rate). Notably, even without access to the instruction datasets, a multimodal instruction set can be successfully poisoned with a very low poisoning rate (0.2%), achieving an attack success rate of over 97%. This paper underscores that even simple traditional backdoor strategies pose a serious threat to LVLMs, necessitating more attention and in-depth research.
Authors: Jr-Jen Chen, Yu-Chien Liao, Hsi-Che Lin, Yu-Chu Yu, Yen-Chun Chen, Yu-Chiang Frank Wang
Abstract: We introduce ReXTime, a benchmark designed to rigorously test AI models' ability to perform temporal reasoning within video events. Specifically, ReXTime focuses on reasoning across time, i.e. human-like understanding when the question and its corresponding answer occur in different video segments. This form of reasoning, requiring advanced understanding of cause-and-effect relationships across video segments, poses significant challenges to even the frontier multimodal large language models. To facilitate this evaluation, we develop an automated pipeline for generating temporal reasoning question-answer pairs, significantly reducing the need for labor-intensive manual annotations. Our benchmark includes 921 carefully vetted validation samples and 2,143 test samples, each manually curated for accuracy and relevance. Evaluation results show that while frontier large language models outperform academic models, they still lag behind human performance by a significant 14.3% accuracy gap. Additionally, our pipeline creates a training dataset of 9,695 machine generated samples without manual effort, which empirical studies suggest can enhance the across-time reasoning via fine-tuning.
Authors: Rikhil Seshadri, Jayant Siva, Angelica Bartholomew, Clara Goebel, Gabriel Wallerstein-King, Beatriz L\'opez Morato, Nicholas Heller, Jason Scovell, Rebecca Campbell, Andrew Wood, Michal Ozery-Flato, Vesna Barros, Maria Gabrani, Michal Rosen-Zvi, Resha Tejpaul, Vidhyalakshmi Ramesh, Nikolaos Papanikolopoulos, Subodh Regmi, Ryan Ward, Robert Abouassaly, Steven C. Campbell, Erick Remer, Christopher Weight
Abstract: Kidney cancer is a global health concern, and accurate assessment of patient frailty is crucial for optimizing surgical outcomes. This paper introduces AI Age Discrepancy, a novel metric derived from machine learning analysis of preoperative abdominal CT scans, as a potential indicator of frailty and postoperative risk in kidney cancer patients. This retrospective study of 599 patients from the 2023 Kidney Tumor Segmentation (KiTS) challenge dataset found that a higher AI Age Discrepancy is significantly associated with longer hospital stays and lower overall survival rates, independent of established factors. This suggests that AI Age Discrepancy may provide valuable insights into patient frailty and could thus inform clinical decision-making in kidney cancer treatment.
Authors: Wenqiang Zu, Shenghao Xie, Qing Zhao, Guoqi Li, Lei Ma
Abstract: Foundation models pre-trained on large-scale data have been widely witnessed to achieve success in various natural imaging downstream tasks. Parameter-efficient fine-tuning (PEFT) methods aim to adapt foundation models to new domains by updating only a small portion of parameters in order to reduce computational overhead. However, the effectiveness of these PEFT methods, especially in cross-domain few-shot scenarios, e.g., medical image analysis, has not been fully explored. In this work, we facilitate the study of the performance of PEFT when adapting foundation models to medical image classification tasks. Furthermore, to alleviate the limitations of prompt introducing ways and approximation capabilities on Transformer architectures of mainstream prompt tuning methods, we propose the Embedded Prompt Tuning (EPT) method by embedding prompt tokens into the expanded channels. We also find that there are anomalies in the feature space distribution of foundation models during pre-training process, and prompt tuning can help mitigate this negative impact. To explain this phenomenon, we also introduce a novel perspective to understand prompt tuning: Prompt tuning is a distribution calibrator. And we support it by analyzing patch-wise scaling and feature separation operations contained in EPT. Our experiments show that EPT outperforms several state-of-the-art fine-tuning methods by a significant margin on few-shot medical image classification tasks, and completes the fine-tuning process within highly competitive time, indicating EPT is an effective PEFT method. The source code is available at github.com/zuwenqiang/EPT.
Authors: Wenhan Yang, Zixuan Hu, Lilang Lin, Jiaying Liu, Ling-Yu Duan
Abstract: Coding, which targets compressing and reconstructing data, and intelligence, often regarded at an abstract computational level as being centered around model learning and prediction, interweave recently to give birth to a series of significant progress. The recent trends demonstrate the potential homogeneity of these two fields, especially when deep-learning models aid these two categories for better probability modeling. For better understanding and describing from a unified perspective, inspired by the basic generally recognized principles in cognitive psychology, we formulate a novel problem of Coding for Intelligence from the category theory view. Based on the three axioms: existence of ideal coding, existence of practical coding, and compactness promoting generalization, we derive a general framework to understand existing methodologies, namely that, coding captures the intrinsic relationships of objects as much as possible, while ignoring information irrelevant to downstream tasks. This framework helps identify the challenges and essential elements in solving the specific derived Minimal Description Length (MDL) optimization problem from a broader range, providing opportunities to build a more intelligent system for handling multiple tasks/applications with coding ideas/tools. Centering on those elements, we systematically review recent processes of towards optimizing the MDL problem in more comprehensive ways from data, model, and task perspectives, and reveal their impacts on the potential CfI technical routes. After that, we also present new technique paths to fulfill CfI and provide potential solutions with preliminary experimental evidence. Last, further directions and remaining issues are discussed as well. The discussion shows our theory can reveal many phenomena and insights about large foundation models, which mutually corroborate with recent practices in feature learning.
Authors: Like Xin, Wanqi Yang, Lei Wang, Ming Yang
Abstract: In this paper, we address the challenging problem of unpaired multi-view clustering (UMC), aiming to perform effective joint clustering using unpaired observed samples across multiple views. Commonly, traditional incomplete multi-view clustering (IMC) methods often depend on paired samples to capture complementary information between views. However, the strategy becomes impractical in UMC due to the absence of paired samples. Although some researchers have attempted to tackle the issue by preserving consistent cluster structures across views, they frequently neglect the confidence of these cluster structures, especially for boundary samples and uncertain cluster structures during the initial training. Therefore, we propose a method called Multi-level Reliable Guidance for UMC (MRG-UMC), which leverages multi-level clustering to aid in learning a trustworthy cluster structure across inner-view, cross-view, and common-view, respectively. Specifically, within each view, multi-level clustering fosters a trustworthy cluster structure across different levels and reduces clustering error. In cross-view learning, reliable view guidance enhances the confidence of the cluster structures in other views. Similarly, within the multi-level framework, the incorporation of a common view aids in aligning different views, thereby reducing the clustering error and uncertainty of cluster structure. Finally, as evidenced by extensive experiments, our method for UMC demonstrates significant efficiency improvements compared to 20 state-of-the-art methods.
Authors: Hongda Liu, Longguang Wang, Ye Zhang, Kaiwen Xue, Shunbo Zhou, Yulan Guo
Abstract: The performance of image super-resolution relies heavily on the accuracy of degradation information, especially under blind settings. Due to absence of true degradation models in real-world scenarios, previous methods learn distinct representations by distinguishing different degradations in a batch. However, the most significant degradation differences may provide shortcuts for the learning of representations such that subtle difference may be discarded. In this paper, we propose an alternative to learn degradation representations through reproducing degraded low-resolution (LR) images. By guiding the degrader to reconstruct input LR images, full degradation information can be encoded into the representations. In addition, we develop an energy distance loss to facilitate the learning of the degradation representations by introducing a bounded constraint. Experiments show that our representations can extract accurate and highly robust degradation information. Moreover, evaluations on both synthetic and real images demonstrate that our ReDSR achieves state-of-the-art performance for the blind SR tasks.
Authors: Chenming Zhu, Tai Wang, Wenwei Zhang, Kai Chen, Xihui Liu
Abstract: Although great progress has been made in 3D visual grounding, current models still rely on explicit textual descriptions for grounding and lack the ability to reason human intentions from implicit instructions. We propose a new task called 3D reasoning grounding and introduce a new benchmark ScanReason which provides over 10K question-answer-location pairs from five reasoning types that require the synerization of reasoning and grounding. We further design our approach, ReGround3D, composed of the visual-centric reasoning module empowered by Multi-modal Large Language Model (MLLM) and the 3D grounding module to obtain accurate object locations by looking back to the enhanced geometry and fine-grained details from the 3D scenes. A chain-of-grounding mechanism is proposed to further boost the performance with interleaved reasoning and grounding steps during inference. Extensive experiments on the proposed benchmark validate the effectiveness of our proposed approach.
Authors: Xiangyang Zhu, Kede Ma, Wufeng Xue
Abstract: Predicting cardiac indices has long been a focal point in the medical imaging community. While various deep learning models have demonstrated success in quantifying cardiac indices, they remain susceptible to mild input perturbations, e.g., spatial transformations, image distortions, and adversarial attacks. This vulnerability undermines confidence in using learning-based automated systems for diagnosing cardiovascular diseases. In this work, we describe a simple yet effective method to learn robust models for left ventricle (LV) quantification, encompassing cavity and myocardium areas, directional dimensions, and regional wall thicknesses. Our success hinges on employing the biologically inspired steerable pyramid transform (SPT) for fixed front-end processing, which offers three main benefits. First, the basis functions of SPT align with the anatomical structure of LV and the geometric features of the measured indices. Second, SPT facilitates weight sharing across different orientations as a form of parameter regularization and naturally captures the scale variations of LV. Third, the residual highpass subband can be conveniently discarded, promoting robust feature learning. Extensive experiments on the Cardiac-Dig benchmark show that our SPT-augmented model not only achieves reasonable prediction accuracy compared to state-of-the-art methods, but also exhibits significantly improved robustness against input perturbations.
Authors: Kuiyuan Zhang, Zhongyun Hua, Yuanman Li, Yushu Zhang, Yicong Zhou
Abstract: Many service computing applications require real-time dataset collection from multiple devices, necessitating efficient sampling techniques to reduce bandwidth and storage pressure. Compressive sensing (CS) has found wide-ranging applications in image acquisition and reconstruction. Recently, numerous deep-learning methods have been introduced for CS tasks. However, the accurate reconstruction of images from measurements remains a significant challenge, especially at low sampling rates. In this paper, we propose Uformer-ICS as a novel U-shaped transformer for image CS tasks by introducing inner characteristics of CS into transformer architecture. To utilize the uneven sparsity distribution of image blocks, we design an adaptive sampling architecture that allocates measurement resources based on the estimated block sparsity, allowing the compressed results to retain maximum information from the original image. Additionally, we introduce a multi-channel projection (MCP) module inspired by traditional CS optimization methods. By integrating the MCP module into the transformer blocks, we construct projection-based transformer blocks, and then form a symmetrical reconstruction model using these blocks and residual convolutional blocks. Therefore, our reconstruction model can simultaneously utilize the local features and long-range dependencies of image, and the prior projection knowledge of CS theory. Experimental results demonstrate its significantly better reconstruction performance than state-of-the-art deep learning-based CS methods.
Authors: Axel Klawonn, Martin Lanser, Janine Weber
Abstract: Deep neural networks (DNNs) and, in particular, convolutional neural networks (CNNs) have brought significant advances in a wide range of modern computer application problems. However, the increasing availability of large amounts of datasets as well as the increasing available computational power of modern computers lead to a steady growth in the complexity and size of DNN and CNN models, respectively, and thus, to longer training times. Hence, various methods and attempts have been developed to accelerate and parallelize the training of complex network architectures. In this work, a novel CNN-DNN architecture is proposed that naturally supports a model parallel training strategy and that is loosely inspired by two-level domain decomposition methods (DDM). First, local CNN models, that is, subnetworks, are defined that operate on overlapping or nonoverlapping parts of the input data, for example, sub-images. The subnetworks can be trained completely in parallel and independently of each other. Each subnetwork then outputs a local decision for the given machine learning problem which is exclusively based on the respective local input data. Subsequently, in a second step, an additional DNN model is trained which evaluates the local decisions of the local subnetworks and generates a final, global decision. In this paper, we apply the proposed architecture to image classification problems using CNNs. Experimental results for different 2D image classification problems are provided as well as a face recognition problem, and a classification problem for 3D computer tomography (CT) scans. Therefore, classical ResNet and VGG architectures are considered. The results show that the proposed approach can significantly accelerate the required training time compared to the global model and, additionally, can also help to improve the accuracy of the underlying classification problem.
Authors: Xiaoxiang Han, Yang Chen, Qiaohong Liu, Yiman Liu, Keyan Chen, Yuanjie Lin, Weikun Zhang
Abstract: Cardiac cine magnetic resonance imaging (MRI) is one of the important means to assess cardiac functions and vascular abnormalities. Mitigating artifacts arising during image reconstruction and accelerating cardiac cine MRI acquisition to obtain high-quality images is important. A novel end-to-end deep learning network is developed to improve cardiac cine MRI reconstruction. First, a U-Net is adopted to obtain the initial reconstructed images in k-space. Further to remove the motion artifacts, the motion-guided deformable alignment (MGDA) module with second-order bidirectional propagation is introduced to align the adjacent cine MRI frames by maximizing spatial-temporal information to alleviate motion artifacts. Finally, the multi-resolution fusion (MRF) module is designed to correct the blur and artifacts generated from alignment operation and obtain the last high-quality reconstructed cardiac images. At an 8$\times$ acceleration rate, the numerical measurements on the ACDC dataset are structural similarity index (SSIM) of 78.40%$\pm$.57%, peak signal-to-noise ratio (PSNR) of 30.46$\pm$1.22dB, and normalized mean squared error (NMSE) of 0.0468$\pm$0.0075. On the ACMRI dataset, the results are SSIM of 87.65%$\pm$4.20%, PSNR of 30.04$\pm$1.18dB, and NMSE of 0.0473$\pm$0.0072. The proposed method exhibits high-quality results with richer details and fewer artifacts for cardiac cine MRI reconstruction on different accelerations.
Authors: Gilles Van De Vyver, Sarina Thomas, Guy Ben-Yosef, Sindre Hellum Olaisen, H\r{a}vard Dalen, Lasse L{\o}vstakken, Erik Smistad
Abstract: Fully automatic cardiac segmentation can be a fast and reproducible method to extract clinical measurements from an echocardiography examination. The U-Net architecture is the current state-of-the-art deep learning architecture for medical segmentation and can segment cardiac structures in real-time with average errors comparable to inter-observer variability. However, this architecture still generates large outliers that are often anatomically incorrect. This work uses the concept of graph convolutional neural networks that predict the contour points of the structures of interest instead of labeling each pixel. We propose a graph architecture that uses two convolutional rings based on cardiac anatomy and show that this eliminates anatomical incorrect multi-structure segmentations on the publicly available CAMUS dataset. Additionally, this work contributes with an ablation study on the graph convolutional architecture and an evaluation of clinical measurements on the clinical HUNT4 dataset. Finally, we propose to use the inter-model agreement of the U-Net and the graph network as a predictor of both the input and segmentation quality. We show this predictor can detect out-of-distribution and unsuitable input images in real-time. Source code is available online: https://github.com/gillesvntnu/GCN_multistructure
Authors: David Kupeev, Eyal Nitcany
Abstract: Every day, humans perceive objects and communicate these perceptions through various channels. In this paper, we present a computational model designed to track and simulate the perception of objects, as well as their representations as conveyed in communication. We delineate two fundamental components of our internal representation, termed "observed" and "seen", which we correlate with established concepts in computer vision, namely encoding and decoding. These components are integrated into semiotic networks, which simulate perceptual inference of object perception and human communication. Our model of object perception by a person allows us to define object perception by {\em a network}. We demonstrate this with an example of an image baseline classifier by constructing a new network that includes the baseline classifier and an additional layer. This layer produces the images "perceived" by the entire network, transforming it into a perceptualized image classifier. This facilitates visualization of the acquired network. Within our network, the image representations become more efficient for classification tasks when they are assembled and randomized. In our experiments, the perceptualized network outperformed the baseline classifier on MNIST training databases consisting of a restricted number of images. Our model is not limited to persons and can be applied to any system featuring a loop involving the processing from "internal" to "external" representations.
Authors: Hejie Ying, Mengmeng Song, Yaohong Tang, Shungen Xiao, Zimin Xiao
Abstract: Deep neural networks (DNN) have achieved remarkable success in various fields, including computer vision and natural language processing. However, training an effective DNN model still poses challenges. This paper aims to propose a method to optimize the training effectiveness of DNN, with the goal of improving model performance. Firstly, based on the observation that the DNN parameters change in certain laws during training process, the potential of parameter prediction for improving model training efficiency and performance is discovered. Secondly, considering the magnitude of DNN model parameters, hardware limitations and characteristics of Stochastic Gradient Descent (SGD) for noise tolerance, a Parameter Linear Prediction (PLP) method is exploit to perform DNN parameter prediction. Finally, validations are carried out on some representative backbones. Experiment results show that compare to the normal training ways, under the same training conditions and epochs, by employing proposed PLP method, the optimal model is able to obtain average about 1% accuracy improvement and 0.01 top-1/top-5 error reduction for Vgg16, Resnet18 and GoogLeNet based on CIFAR-100 dataset, which shown the effectiveness of the proposed method on different DNN structures, and validated its capacity in enhancing DNN training efficiency and performance.
Authors: Gia Minh Hoang, Youngjoo Lee, Jae Gwan Kim
Abstract: Alzheimer's disease is one of the most common types of neurodegenerative disease, characterized by the accumulation of amyloid-beta plaque and tau tangles. Recently, deep learning approaches have shown promise in Alzheimer's disease diagnosis. In this study, we propose a reproducible model that utilizes a 3D convolutional neural network with a dual attention module for Alzheimer's disease classification. We trained the model in the ADNI database and verified the generalizability of our method in two independent datasets (AIBL and OASIS1). Our method achieved state-of-the-art classification performance, with an accuracy of 91.94% for MCI progression classification and 96.30% for Alzheimer's disease classification on the ADNI dataset. Furthermore, the model demonstrated good generalizability, achieving an accuracy of 86.37% on the AIBL dataset and 83.42% on the OASIS1 dataset. These results indicate that our proposed approach has competitive performance and generalizability when compared to recent studies in the field.
Authors: Emma Cramer, Jonas Reiher, Sebastian Trimpe
Abstract: Reinforcement learning (RL) for robot control typically requires a detailed representation of the environment state, including information about task-relevant objects not directly measurable. Keypoint detectors, such as spatial autoencoders (SAEs), are a common approach to extracting a low-dimensional representation from high-dimensional image data. SAEs aim at spatial features such as object positions, which are often useful representations in robotic RL. However, whether an SAE is actually able to track objects in the scene and thus yields a spatial state representation well suited for RL tasks has rarely been examined due to a lack of established metrics. In this paper, we propose to assess the performance of an SAE instance by measuring how well keypoints track ground truth objects in images. We present a computationally lightweight metric and use it to evaluate common baseline SAE architectures on image data from a simulated robot task. We find that common SAEs differ substantially in their spatial extraction capability. Furthermore, we validate that SAEs that perform well in our metric achieve superior performance when used in downstream RL. Thus, our metric is an effective and lightweight indicator of RL performance before executing expensive RL training. Building on these insights, we identify three key modifications of SAE architectures to improve tracking performance.
Authors: Yue Pan, Xingguang Zhong, Louis Wiesmann, Thorbj\"orn Posewsky, Jens Behley, Cyrill Stachniss
Abstract: Accurate and robust localization and mapping are essential components for most autonomous robots. In this paper, we propose a SLAM system for building globally consistent maps, called PIN-SLAM, that is based on an elastic and compact point-based implicit neural map representation. Taking range measurements as input, our approach alternates between incremental learning of the local implicit signed distance field and the pose estimation given the current local map using a correspondence-free, point-to-implicit model registration. Our implicit map is based on sparse optimizable neural points, which are inherently elastic and deformable with the global pose adjustment when closing a loop. Loops are also detected using the neural point features. Extensive experiments validate that PIN-SLAM is robust to various environments and versatile to different range sensors such as LiDAR and RGB-D cameras. PIN-SLAM achieves pose estimation accuracy better or on par with the state-of-the-art LiDAR odometry or SLAM systems and outperforms the recent neural implicit SLAM approaches while maintaining a more consistent, and highly compact implicit map that can be reconstructed as accurate and complete meshes. Finally, thanks to the voxel hashing for efficient neural points indexing and the fast implicit map-based registration without closest point association, PIN-SLAM can run at the sensor frame rate on a moderate GPU. Codes will be available at: https://github.com/PRBonn/PIN_SLAM.
Authors: Michael Ahn, Debidatta Dwibedi, Chelsea Finn, Montse Gonzalez Arenas, Keerthana Gopalakrishnan, Karol Hausman, Brian Ichter, Alex Irpan, Nikhil Joshi, Ryan Julian, Sean Kirmani, Isabel Leal, Edward Lee, Sergey Levine, Yao Lu, Isabel Leal, Sharath Maddineni, Kanishka Rao, Dorsa Sadigh, Pannag Sanketi, Pierre Sermanet, Quan Vuong, Stefan Welker, Fei Xia, Ted Xiao, Peng Xu, Steve Xu, Zhuo Xu
Abstract: Foundation models that incorporate language, vision, and more recently actions have revolutionized the ability to harness internet scale data to reason about useful tasks. However, one of the key challenges of training embodied foundation models is the lack of data grounded in the physical world. In this paper, we propose AutoRT, a system that leverages existing foundation models to scale up the deployment of operational robots in completely unseen scenarios with minimal human supervision. AutoRT leverages vision-language models (VLMs) for scene understanding and grounding, and further uses large language models (LLMs) for proposing diverse and novel instructions to be performed by a fleet of robots. Guiding data collection by tapping into the knowledge of foundation models enables AutoRT to effectively reason about autonomy tradeoffs and safety while significantly scaling up data collection for robot learning. We demonstrate AutoRT proposing instructions to over 20 robots across multiple buildings and collecting 77k real robot episodes via both teleoperation and autonomous robot policies. We experimentally show that such "in-the-wild" data collected by AutoRT is significantly more diverse, and that AutoRT's use of LLMs allows for instruction following data collection robots that can align to human preferences.
Authors: Livia Rodrigues, Martina Bocchetta, Oula Puonti, Douglas Greve, Ana Carolina Londe, Marcondes Fran\c{c}a, Simone Appenzeller, Juan Eugenio Iglesias, Leticia Rittner
Abstract: The hypothalamus is a small structure located in the center of the brain and is involved in significant functions such as sleeping, temperature, and appetite control. Various neurological disorders are also associated with hypothalamic abnormalities. Automated image analysis of this structure from brain MRI is thus highly desirable to study the hypothalamus in vivo. However, most automated segmentation tools currently available focus exclusively on T1w images. In this study, we introduce H-SynEx, a machine learning method for automated segmentation of hypothalamic subregions that generalizes across different MRI sequences and resolutions without retraining. H-synEx was trained with synthetic images built from label maps derived from ultra-high resolution ex vivo MRI scans, which enables finer-grained manual segmentation when compared with 1mm isometric in vivo images. We validated our method using Dice Coefficient (DSC) and Average Hausdorff distance (AVD) across in vivo images from six different datasets with six different MRI sequences (T1, T2, proton density, quantitative T1, fractional anisotrophy, and FLAIR). Statistical analysis compared hypothalamic subregion volumes in controls, Alzheimer's disease (AD), and behavioral variant frontotemporal dementia (bvFTD) subjects using the Area Under the Receiving Operating Characteristic curve (AUROC) and Wilcoxon rank sum test. Our results show that H-SynEx successfully leverages information from ultra-high resolution scans to segment in vivo from different MRI sequences. Our automated segmentation was able to discriminate controls versus Alzheimer's Disease patients on FLAIR images with 5mm spacing. H-SynEx is openly available at https://github.com/liviamarodrigues/hsynex.
Authors: Matt Y Cheung, Tucker J Netherton, Laurence E Court, Ashok Veeraraghavan, Guha Balakrishnan
Abstract: Recent advancements in machine learning have led to the development of novel medical imaging systems and algorithms that address ill-posed problems. Assessing their trustworthiness and understanding how to deploy them safely at test time remains an important and open problem. In this work, we propose using conformal prediction to compute valid and distribution-free bounds on downstream metrics given reconstructions generated by one algorithm, and retrieve upper/lower bounds and inlier/outlier reconstructions according to the adjusted bounds. Our work offers 1) test time image reconstruction evaluation without ground truth, 2) downstream performance guarantees, 3) meaningful upper/lower bound reconstructions, and 4) meaningful statistical inliers/outlier reconstructions. We demonstrate our method on post-mastectomy radiotherapy planning using 3D breast CT reconstructions, and show 1) that metric-guided bounds have valid coverage for downstream metrics while conventional pixel-wise bounds do not and 2) anatomical differences of upper/lower bounds between metric-guided and pixel-wise methods. Our work paves way for more meaningful and trustworthy test-time evaluation of medical image reconstructions. Code available at https://github.com/matthewyccheung/conformal-metric
Authors: Albert Yu, Adeline Foote, Raymond Mooney, Roberto Mart\'in-Mart\'in
Abstract: The main challenge in learning image-conditioned robotic policies is acquiring a visual representation conducive to low-level control. Due to the high dimensionality of the image space, learning a good visual representation requires a considerable amount of visual data. However, when learning in the real world, data is expensive. Sim2Real is a promising paradigm for overcoming data scarcity in the real-world target domain by using a simulator to collect large amounts of cheap data closely related to the target task. However, it is difficult to transfer an image-conditioned policy from sim to real when the domains are very visually dissimilar. To bridge the sim2real visual gap, we propose using natural language descriptions of images as a unifying signal across domains that captures the underlying task-relevant semantics. Our key insight is that if two image observations from different domains are labeled with similar language, the policy should predict similar action distributions for both images. We demonstrate that training the image encoder to predict the language description or the distance between descriptions of a sim or real image serves as a useful, data-efficient pretraining step that helps learn a domain-invariant image representation. We can then use this image encoder as the backbone of an IL policy trained simultaneously on a large amount of simulated and a handful of real demonstrations. Our approach outperforms widely used prior sim2real methods and strong vision-language pretraining baselines like CLIP and R3M by 25 to 40%. See additional videos and materials at https://robin-lab.cs.utexas.edu/lang4sim2real/.
Authors: Chao-Hui Huang
Abstract: Recently, various technologies have been introduced into digital pathology, including artificial intelligence (AI) driven methods, in both areas of pathological whole slide image (WSI) analysis and spatial transcriptomics (ST) analysis. AI-driven WSI analysis utilizes the power of deep learning (DL), expands the field of view for histopathological image analysis. On the other hand, ST bridges the gap between tissue spatial analysis and biological signals, offering the possibility to understand the spatial biology. However, a major bottleneck in DL-based WSI analysis is the preparation of training patterns, as hematoxylin & eosin (H&E) staining does not provide direct biological evidence, such as gene expression, for determining the category of a biological component. On the other hand, as of now, the resolution in ST is far beyond that of WSI, resulting the challenge of further spatial analysis. Although various WSI analysis tools, including QuPath, have cited the use of WSI analysis tools in the context of ST analysis, its usage is primarily focused on initial image analysis, with other tools being utilized for more detailed transcriptomic analysis. As a result, the information hidden beneath WSI has not yet been fully utilized to support ST analysis. To bridge this gap, we introduce QuST, a QuPath extension designed to bridge the gap between H&E WSI and ST analyzing tasks. In this paper, we highlight the importance of integrating DL-based WSI analysis and ST analysis in understanding disease biology and the challenges in integrating these modalities due to differences in data formats and analytical methods. The QuST source code is hosted on GitHub and documentation is available at (https://github.com/huangch/qust).
Authors: Hamed Karimi, Reza Samavi
Abstract: In this paper, we propose Evidential Conformal Prediction (ECP) method for image classifiers to generate the conformal prediction sets. Our method is designed based on a non-conformity score function that has its roots in Evidential Deep Learning (EDL) as a method of quantifying model (epistemic) uncertainty in DNN classifiers. We use evidence that are derived from the logit values of target labels to compute the components of our non-conformity score function: the heuristic notion of uncertainty in CP, uncertainty surprisal, and expected utility. Our extensive experimental evaluation demonstrates that ECP outperforms three state-of-the-art methods for generating CP sets, in terms of their set sizes and adaptivity while maintaining the coverage of true labels.
Authors: Chao Hui Huang
Abstract: In this paper, we introduce QuST-LLM, an innovative extension of QuPath that utilizes the capabilities of large language models (LLMs) to analyze and interpret spatial transcriptomics (ST) data. In addition to simplifying the intricate and high-dimensional nature of ST data by offering a comprehensive workflow that includes data loading, region selection, gene expression analysis, and functional annotation, QuST-LLM employs LLMs to transform complex ST data into understandable and detailed biological narratives based on gene ontology annotations, thereby significantly improving the interpretability of ST data. Consequently, users can interact with their own ST data using natural language. Hence, QuST-LLM provides researchers with a potent functionality to unravel the spatial and functional complexities of tissues, fostering novel insights and advancements in biomedical research. QuST-LLM is a part of QuST project. The source code is hosted on GitHub and documentation is available at (https://github.com/huangch/qust).
Authors: Chen Liu, Ke Xu, Liangbo L. Shen, Guillaume Huguet, Zilong Wang, Alexander Tong, Danilo Bzdok, Jay Stewart, Jay C. Wang, Lucian V. Del Priore, Smita Krishnaswamy
Abstract: The forecasting of disease progression from images is a holy grail for clinical decision making. However, this task is complicated by the inherent high dimensionality, temporal sparsity and sampling irregularity in longitudinal image acquisitions. Existing methods often rely on extracting hand-crafted features and performing time-series analysis in this vector space, leading to a loss of rich spatial information within the images. To overcome these challenges, we introduce ImageFlowNet, a novel framework that learns latent-space flow fields that evolve multiscale representations in joint embedding spaces using neural ODEs and SDEs to model disease progression in the image domain. Notably, ImageFlowNet learns multiscale joint representation spaces by combining cohorts of patients together so that information can be transferred between the patient samples. The dynamics then provide plausible trajectories of progression, with the SDE providing alternative trajectories from the same starting point. We provide theoretical insights that support our formulation of ODEs, and motivate our regularizations involving high-level visual features, latent space organization, and trajectory smoothness. We then demonstrate ImageFlowNet's effectiveness through empirical evaluations on three longitudinal medical image datasets depicting progression in retinal geographic atrophy, multiple sclerosis, and glioblastoma.
Authors: Georgios Tziafas, Hamidreza Kasaei
Abstract: The ability to grasp objects in-the-wild from open-ended language instructions constitutes a fundamental challenge in robotics. An open-world grasping system should be able to combine high-level contextual with low-level physical-geometric reasoning in order to be applicable in arbitrary scenarios. Recent works exploit the web-scale knowledge inherent in large language models (LLMs) to plan and reason in robotic context, but rely on external vision and action models to ground such knowledge into the environment and parameterize actuation. This setup suffers from two major bottlenecks: a) the LLM's reasoning capacity is constrained by the quality of visual grounding, and b) LLMs do not contain low-level spatial understanding of the world, which is essential for grasping in contact-rich scenarios. In this work we demonstrate that modern vision-language models (VLMs) are capable of tackling such limitations, as they are implicitly grounded and can jointly reason about semantics and geometry. We propose OWG, an open-world grasping pipeline that combines VLMs with segmentation and grasp synthesis models to unlock grounded world understanding in three stages: open-ended referring segmentation, grounded grasp planning and grasp ranking via contact reasoning, all of which can be applied zero-shot via suitable visual prompting mechanisms. We conduct extensive evaluation in cluttered indoor scene datasets to showcase OWG's robustness in grounding from open-ended language, as well as open-world robotic grasping experiments in both simulation and hardware that demonstrate superior performance compared to previous supervised and zero-shot LLM-based methods.
Authors: Youngmin Seo, Jinha Kim, Unsang Park
Abstract: We propose the Swish-T family, an enhancement of the existing non-monotonic activation function Swish. Swish-T is defined by adding a Tanh bias to the original Swish function. This modification creates a family of Swish-T variants, each designed to excel in different tasks, showcasing specific advantages depending on the application context. The Tanh bias allows for broader acceptance of negative values during initial training stages, offering a smoother non-monotonic curve than the original Swish. We ultimately propose the Swish-T$_{\textbf{C}}$ function, while Swish-T and Swish-T$_{\textbf{B}}$, byproducts of Swish-T$_{\textbf{C}}$, also demonstrate satisfactory performance. Furthermore, our ablation study shows that using Swish-T$_{\textbf{C}}$ as a non-parametric function can still achieve high performance. The superiority of the Swish-T family has been empirically demonstrated across various models and benchmark datasets, including MNIST, Fashion MNIST, SVHN, CIFAR-10, and CIFAR-100. The code is publicly available at "https://github.com/ictseoyoungmin/Swish-T-pytorch".
Authors: Boyuan Chen, Diego Marti Monso, Yilun Du, Max Simchowitz, Russ Tedrake, Vincent Sitzmann
Abstract: This paper presents Diffusion Forcing, a new training paradigm where a diffusion model is trained to denoise a set of tokens with independent per-token noise levels. We apply Diffusion Forcing to sequence generative modeling by training a causal next-token prediction model to generate one or several future tokens without fully diffusing past ones. Our approach is shown to combine the strengths of next-token prediction models, such as variable-length generation, with the strengths of full-sequence diffusion models, such as the ability to guide sampling to desirable trajectories. Our method offers a range of additional capabilities, such as (1) rolling-out sequences of continuous tokens, such as video, with lengths past the training horizon, where baselines diverge and (2) new sampling and guiding schemes that uniquely profit from Diffusion Forcing's variable-horizon and causal architecture, and which lead to marked performance gains in decision-making and planning tasks. In addition to its empirical success, our method is proven to optimize a variational lower bound on the likelihoods of all subsequences of tokens drawn from the true joint distribution. Project website: https://boyuan.space/diffusion-forcing/
Authors: Manuel Faysse, Hugues Sibille, Tony Wu, Bilel Omrani, Gautier Viaud, C\'eline Hudelot, Pierre Colombo
Abstract: Documents are visually rich structures that convey information through text, as well as tables, figures, page layouts, or fonts. While modern document retrieval systems exhibit strong performance on query-to-text matching, they struggle to exploit visual cues efficiently, hindering their performance on practical document retrieval applications such as Retrieval Augmented Generation. To benchmark current systems on visually rich document retrieval, we introduce the Visual Document Retrieval Benchmark ViDoRe, composed of various page-level retrieving tasks spanning multiple domains, languages, and settings. The inherent shortcomings of modern systems motivate the introduction of a new retrieval model architecture, ColPali, which leverages the document understanding capabilities of recent Vision Language Models to produce high-quality contextualized embeddings solely from images of document pages. Combined with a late interaction matching mechanism, ColPali largely outperforms modern document retrieval pipelines while being drastically faster and end-to-end trainable.
Authors: Tianrun Chen, Chaotao Ding, Lanyun Zhu, Tao Xu, Deyi Ji, Yan Wang, Ying Zang, Zejian Li
Abstract: Convolutional Neural Networks (CNNs) and Vision Transformers (ViT) have been pivotal in biomedical image segmentation, yet their ability to manage long-range dependencies remains constrained by inherent locality and computational overhead. To overcome these challenges, in this technical report, we first propose xLSTM-UNet, a UNet structured deep learning neural network that leverages Vision-LSTM (xLSTM) as its backbone for medical image segmentation. xLSTM is a recently proposed as the successor of Long Short-Term Memory (LSTM) networks and have demonstrated superior performance compared to Transformers and State Space Models (SSMs) like Mamba in Neural Language Processing (NLP) and image classification (as demonstrated in Vision-LSTM, or ViL implementation). Here, xLSTM-UNet we designed extend the success in biomedical image segmentation domain. By integrating the local feature extraction strengths of convolutional layers with the long-range dependency capturing abilities of xLSTM, xLSTM-UNet offers a robust solution for comprehensive image analysis. We validate the efficacy of xLSTM-UNet through experiments. Our findings demonstrate that xLSTM-UNet consistently surpasses the performance of leading CNN-based, Transformer-based, and Mamba-based segmentation networks in multiple datasets in biomedical segmentation including organs in abdomen MRI, instruments in endoscopic images, and cells in microscopic images. With comprehensive experiments performed, this technical report highlights the potential of xLSTM-based architectures in advancing biomedical image analysis in both 2D and 3D. The code, models, and datasets are publicly available at http://tianrun-chen.github.io/xLSTM-UNet/