Authors: Abdullah Nazhat Abdullah, Tarkan Aydin
Abstract: The Attention mechanism is the main component of the Transformer architecture, and since its introduction, it has led to significant advancements in Deep Learning that span many domains and multiple tasks. The Attention Mechanism was utilized in Computer Vision as the Vision Transformer ViT, and its usage has expanded into many tasks in the vision domain, such as classification, segmentation, object detection, and image generation. While this mechanism is very expressive and capable, it comes with the drawback of being computationally expensive and requiring datasets of considerable size for effective optimization. To address these shortcomings, many designs have been proposed in the literature to reduce the computational burden and alleviate the data size requirements. Examples of such attempts in the vision domain are the MLP-Mixer, the Conv-Mixer, the Perciver-IO, and many more. This paper introduces a new computational block as an alternative to the standard ViT block that reduces the compute burdens by replacing the normal Attention layers with a Network in Network structure that enhances the static approach of the MLP Mixer with a dynamic system of learning an element-wise gating function by a token mixing process. Extensive experimentation shows that the proposed design provides better performance than the baseline architectures on multiple datasets applied in the image classification task of the vision domain.
Authors: Camilo Calixto, Camilo Jaimes, Matheus D. Soldatelli, Simon K. Warfield, Ali Gholipour, Davood Karimi
Abstract: Diffusion-weighted Magnetic Resonance Imaging (dMRI) is increasingly used to study the fetal brain in utero. An important computation enabled by dMRI is streamline tractography, which has unique applications such as tract-specific analysis of the brain white matter and structural connectivity assessment. However, due to the low fetal dMRI data quality and the challenging nature of tractography, existing methods tend to produce highly inaccurate results. They generate many false streamlines while failing to reconstruct streamlines that constitute the major white matter tracts. In this paper, we advocate for anatomically constrained tractography based on an accurate segmentation of the fetal brain tissue directly in the dMRI space. We develop a deep learning method to compute the segmentation automatically. Experiments on independent test data show that this method can accurately segment the fetal brain tissue and drastically improve tractography results. It enables the reconstruction of highly curved tracts such as optic radiations. Importantly, our method infers the tissue segmentation and streamline propagation direction from a diffusion tensor fit to the dMRI data, making it applicable to routine fetal dMRI scans. The proposed method can lead to significant improvements in the accuracy and reproducibility of quantitative assessment of the fetal brain with dMRI.
Authors: Mahmoud Afifi, Zhenhua Hu, Liang Liang
Abstract: High dynamic range (HDR) imaging involves capturing a series of frames of the same scene, each with different exposure settings, to broaden the dynamic range of light. This can be achieved through burst capturing or using staggered HDR sensors that capture long and short exposures simultaneously in the camera image signal processor (ISP). Within camera ISP pipeline, illuminant estimation is a crucial step aiming to estimate the color of the global illuminant in the scene. This estimation is used in camera ISP white-balance module to remove undesirable color cast in the final image. Despite the multiple frames captured in the HDR pipeline, conventional illuminant estimation methods often rely only on a single frame of the scene. In this paper, we explore leveraging information from frames captured with different exposure times. Specifically, we introduce a simple feature extracted from dual-exposure images to guide illuminant estimators, referred to as the dual-exposure feature (DEF). To validate the efficiency of DEF, we employed two illuminant estimators using the proposed DEF: 1) a multilayer perceptron network (MLP), referred to as exposure-based MLP (EMLP), and 2) a modified version of the convolutional color constancy (CCC) to integrate our DEF, that we call ECCC. Both EMLP and ECCC achieve promising results, in some cases surpassing prior methods that require hundreds of thousands or millions of parameters, with only a few hundred parameters for EMLP and a few thousand parameters for ECCC.
Authors: Iryna Hartsock, Ghulam Rasool
Abstract: Medical vision-language models (VLMs) combine computer vision and natural language processing to analyze visual and textual medical data. Our paper reviews recent advancements in developing VLMs specialized for healthcare, focusing on models designed for medical report generation and visual question answering. We provide background on natural language processing and computer vision, explaining how techniques from both fields are integrated into VLMs to enable learning from multimodal data. Key areas we address include the exploration of medical vision-language datasets, in-depth analyses of architectures and pre-training strategies employed in recent noteworthy medical VLMs, and comprehensive discussion on evaluation metrics for assessing VLMs' performance in medical report generation and visual question answering. We also highlight current challenges and propose future directions, including enhancing clinical validity and addressing patient privacy concerns. Overall, our review summarizes recent progress in developing VLMs to harness multimodal medical data for improved healthcare applications.
Authors: Sahan Ahmad, Gabriel Trahan, Aminul Islam
Abstract: Convolutional Neural Networks (CNNs) have demonstrated outstanding performance in computer vision tasks such as image classification, detection, segmentation, and medical image analysis. In general, an arbitrary number of epochs is used to train such neural networks. In a single epoch, the entire training data -- divided by batch size -- are fed to the network. In practice, validation error with training loss is used to estimate the neural network's generalization, which indicates the optimal learning capacity of the network. Current practice is to stop training when the training loss decreases and the gap between training and validation error increases (i.e., the generalization gap) to avoid overfitting. However, this is a trial-and-error-based approach which raises a critical question: Is it possible to estimate when neural networks stop learning based on training data? This research work introduces a hypothesis that analyzes the data variation across all the layers of a CNN variant to anticipate its near-optimal learning capacity. In the training phase, we use our hypothesis to anticipate the near-optimal learning capacity of a CNN variant without using any validation data. Our hypothesis can be deployed as a plug-and-play to any existing CNN variant without introducing additional trainable parameters to the network. We test our hypothesis on six different CNN variants and three different general image datasets (CIFAR10, CIFAR100, and SVHN). The result based on these CNN variants and datasets shows that our hypothesis saves 58.49\% of computational time (on average) in training. We further conduct our hypothesis on ten medical image datasets and compared with the MedMNIST-V2 benchmark. Based on our experimental result, we save $\approx$ 44.1\% of computational time without losing accuracy against the MedMNIST-V2 benchmark.
Authors: Tom Sander, Yaodong Yu, Maziar Sanjabi, Alain Durmus, Yi Ma, Kamalika Chaudhuri, Chuan Guo
Abstract: Differentially private (DP) machine learning is considered the gold-standard solution for training a model from sensitive data while still preserving privacy. However, a major barrier to achieving this ideal is its sub-optimal privacy-accuracy trade-off, which is particularly visible in DP representation learning. Specifically, it has been shown that under modest privacy budgets, most models learn representations that are not significantly better than hand-crafted features. In this work, we show that effective DP representation learning can be done via image captioning and scaling up to internet-scale multimodal datasets. Through a series of engineering tricks, we successfully train a DP image captioner (DP-Cap) on a 233M subset of LAION-2B from scratch using a reasonable amount of computation, and obtaining unprecedented high-quality image features that can be used in a variety of downstream vision and vision-language tasks. For example, under a privacy budget of $\varepsilon=8$, a linear classifier trained on top of learned DP-Cap features attains 65.8% accuracy on ImageNet-1K, considerably improving the previous SOTA of 56.5%. Our work challenges the prevailing sentiment that high-utility DP representation learning cannot be achieved by training from scratch.
Authors: Amariah Condon, Bailey Buscarino, Eric Moch, William J. Sehnert, Owen Miles, Patrick K. Herring, Peter M. Attia
Abstract: Battery technology is increasingly important for global electrification efforts. However, batteries are highly sensitive to small manufacturing variations that can induce reliability or safety issues. An important technology for battery quality control is computed tomography (CT) scanning, which is widely used for non-destructive 3D inspection across a variety of clinical and industrial applications. Historically, however, the utility of CT scanning for high-volume manufacturing has been limited by its low throughput as well as the difficulty of handling its large file sizes. In this work, we present a dataset of over one thousand CT scans of as-produced commercially available batteries. The dataset spans various chemistries (lithium-ion and sodium-ion) as well as various battery form factors (cylindrical, pouch, and prismatic). We evaluate seven different battery types in total. The manufacturing variability and the presence of battery defects can be observed via this dataset. This dataset may be of interest to scientists and engineers working on battery technology, computer vision, or both.
Authors: Xiaoyu Zhan, Jianxin Yang, Yuanqi Li, Jie Guo, Yanwen Guo, Wenping Wang
Abstract: The field of 3D detailed human mesh reconstruction has made significant progress in recent years. However, current methods still face challenges when used in industrial applications due to unstable results, low-quality meshes, and a lack of UV unwrapping and skinning weights. In this paper, we present SHERT, a novel pipeline that can reconstruct semantic human meshes with textures and high-precision details. SHERT applies semantic- and normal-based sampling between the detailed surface (eg mesh and SDF) and the corresponding SMPL-X model to obtain a partially sampled semantic mesh and then generates the complete semantic mesh by our specifically designed self-supervised completion and refinement networks. Using the complete semantic mesh as a basis, we employ a texture diffusion model to create human textures that are driven by both images and texts. Our reconstructed meshes have stable UV unwrapping, high-quality triangle meshes, and consistent semantic information. The given SMPL-X model provides semantic information and shape priors, allowing SHERT to perform well even with incorrect and incomplete inputs. The semantic information also makes it easy to substitute and animate different body parts such as the face, body, and hands. Quantitative and qualitative experiments demonstrate that SHERT is capable of producing high-fidelity and robust semantic meshes that outperform state-of-the-art methods.
Authors: Aashaka Desai, Maartje De Meulder, Julie A. Hochgesang, Annemarie Kocab, Alex X. Lu
Abstract: Growing research in sign language recognition, generation, and translation AI has been accompanied by calls for ethical development of such technologies. While these works are crucial to helping individual researchers do better, there is a notable lack of discussion of systemic biases or analysis of rhetoric that shape the research questions and methods in the field, especially as it remains dominated by hearing non-signing researchers. Therefore, we conduct a systematic review of 101 recent papers in sign language AI. Our analysis identifies significant biases in the current state of sign language AI research, including an overfocus on addressing perceived communication barriers, a lack of use of representative datasets, use of annotations lacking linguistic foundations, and development of methods that build on flawed models. We take the position that the field lacks meaningful input from Deaf stakeholders, and is instead driven by what decisions are the most convenient or perceived as important to hearing researchers. We end with a call to action: the field must make space for Deaf researchers to lead the conversation in sign language AI.
Authors: Hamid Kazemi, Atoosa Chegini, Jonas Geiping, Soheil Feizi, Tom Goldstein
Abstract: We employ an inversion-based approach to examine CLIP models. Our examination reveals that inverting CLIP models results in the generation of images that exhibit semantic alignment with the specified target prompts. We leverage these inverted images to gain insights into various aspects of CLIP models, such as their ability to blend concepts and inclusion of gender biases. We notably observe instances of NSFW (Not Safe For Work) images during model inversion. This phenomenon occurs even for semantically innocuous prompts, like "a beautiful landscape," as well as for prompts involving the names of celebrities.
Authors: Zhiyuan Chang, Mingyang Li, Junjie Wang, Cheng Li, Qing Wang
Abstract: Visual entailment (VE) is a multimodal reasoning task consisting of image-sentence pairs whereby a promise is defined by an image, and a hypothesis is described by a sentence. The goal is to predict whether the image semantically entails the sentence. VE systems have been widely adopted in many downstream tasks. Metamorphic testing is the commonest technique for AI algorithms, but it poses a significant challenge for VE testing. They either only consider perturbations on single modality which would result in ineffective tests due to the destruction of the relationship of image-text pair, or just conduct shallow perturbations on the inputs which can hardly detect the decision error made by VE systems. Motivated by the fact that objects in the image are the fundamental element for reasoning, we propose VEglue, an object-aligned joint erasing approach for VE systems testing. It first aligns the object regions in the premise and object descriptions in the hypothesis to identify linked and un-linked objects. Then, based on the alignment information, three Metamorphic Relations are designed to jointly erase the objects of the two modalities. We evaluate VEglue on four widely-used VE systems involving two public datasets. Results show that VEglue could detect 11,609 issues on average, which is 194%-2,846% more than the baselines. In addition, VEglue could reach 52.5% Issue Finding Rate (IFR) on average, and significantly outperform the baselines by 17.1%-38.2%. Furthermore, we leverage the tests generated by VEglue to retrain the VE systems, which largely improves model performance (50.8% increase in accuracy) on newly generated tests without sacrificing the accuracy on the original test set.
Authors: Yuelin Zhang, Pengyu Zheng, Wanquan Yan, Chengyu Fang, Shing Shin Cheng
Abstract: Defocus blur is a persistent problem in microscope imaging that poses harm to pathology interpretation and medical intervention in cell microscopy and microscope surgery. To address this problem, a unified framework including multi-pyramid transformer (MPT) and extended frequency contrastive regularization (EFCR) is proposed to tackle two outstanding challenges in microscopy deblur: longer attention span and feature deficiency. The MPT employs an explicit pyramid structure at each network stage that integrates the cross-scale window attention (CSWA), the intra-scale channel attention (ISCA), and the feature-enhancing feed-forward network (FEFN) to capture long-range cross-scale spatial interaction and global channel context. The EFCR addresses the feature deficiency problem by exploring latent deblur signals from different frequency bands. It also enables deblur knowledge transfer to learn cross-domain information from extra data, improving deblur performance for labeled and unlabeled data. Extensive experiments and downstream task validation show the framework achieves state-of-the-art performance across multiple datasets. Project page: https://github.com/PieceZhang/MPT-CataBlur.
Authors: Imad Eddine Toubal, Aditya Avinash, Neil Gordon Alldrin, Jan Dlabal, Wenlei Zhou, Enming Luo, Otilia Stretcu, Hao Xiong, Chun-Ta Lu, Howard Zhou, Ranjay Krishna, Ariel Fuxman, Tom Duerig
Abstract: From content moderation to wildlife conservation, the number of applications that require models to recognize nuanced or subjective visual concepts is growing. Traditionally, developing classifiers for such concepts requires substantial manual effort measured in hours, days, or even months to identify and annotate data needed for training. Even with recently proposed Agile Modeling techniques, which enable rapid bootstrapping of image classifiers, users are still required to spend 30 minutes or more of monotonous, repetitive data labeling just to train a single classifier. Drawing on Fiske's Cognitive Miser theory, we propose a new framework that alleviates manual effort by replacing human labeling with natural language interactions, reducing the total effort required to define a concept by an order of magnitude: from labeling 2,000 images to only 100 plus some natural language interactions. Our framework leverages recent advances in foundation models, both large language models and vision-language models, to carve out the concept space through conversation and by automatically labeling training data points. Most importantly, our framework eliminates the need for crowd-sourced annotations. Moreover, our framework ultimately produces lightweight classification models that are deployable in cost-sensitive scenarios. Across 15 subjective concepts and across 2 public image classification datasets, our trained models outperform traditional Agile Modeling as well as state-of-the-art zero-shot classification models like ALIGN, CLIP, CuPL, and large visual question-answering models like PaLI-X.
Authors: Biqing Qi, Xingquan Chen, Junqi Gao, Jianxing Liu, Ligang Wu, Bowen Zhou
Abstract: Advanced life forms, sustained by the synergistic interaction of neural cognitive mechanisms, continually acquire and transfer knowledge throughout their lifespan. In contrast, contemporary machine learning paradigms exhibit limitations in emulating the facets of continual learning (CL). Nonetheless, the emergence of large language models (LLMs) presents promising avenues for realizing CL via interactions with these models. Drawing on Complementary Learning System theory, this paper presents a novel Interactive Continual Learning (ICL) framework, enabled by collaborative interactions among models of various sizes. Specifically, we assign the ViT model as System1 and multimodal LLM as System2. To enable the memory module to deduce tasks from class information and enhance Set2Set retrieval, we propose the Class-Knowledge-Task Multi-Head Attention (CKT-MHA). Additionally, to improve memory retrieval in System1 through enhanced geometric representation, we introduce the CL-vMF mechanism, based on the von Mises-Fisher (vMF) distribution. Meanwhile, we introduce the von Mises-Fisher Outlier Detection and Interaction (vMF-ODI) strategy to identify hard examples, thus enhancing collaboration between System1 and System2 for complex reasoning realization. Comprehensive evaluation of our proposed ICL demonstrates significant resistance to forgetting and superior performance relative to existing methods.
Authors: Yu Chen, Liyan Ma, Liping Jing, Jian Yu
Abstract: Humans can easily distinguish the known and unknown categories and can recognize the unknown object by learning it once instead of repeating it many times without forgetting the learned object. Hence, we aim to make deep learning models simulate the way people learn. We refer to such a learning manner as OnLine Open World Object Detection(OLOWOD). Existing OWOD approaches pay more attention to the identification of unknown categories, while the incremental learning part is also very important. Besides, some neuroscience research shows that specific noises allow the brain to form new connections and neural pathways which may improve learning speed and efficiency. In this paper, we take the dual-level information of old samples as perturbations on new samples to make the model good at learning new knowledge without forgetting the old knowledge. Therefore, we propose a simple plug-and-play method, called Brain-inspired Streaming Dual-level Perturbations(BSDP), to solve the OLOWOD problem. Specifically, (1) we first calculate the prototypes of previous categories and use the distance between samples and the prototypes as the sample selecting strategy to choose old samples for replay; (2) then take the prototypes as the streaming feature-level perturbations of new samples, so as to improve the plasticity of the model through revisiting the old knowledge; (3) and also use the distribution of the features of the old category samples to generate adversarial data in the form of streams as the data-level perturbations to enhance the robustness of the model to new categories. We empirically evaluate BSDP on PASCAL VOC and MS-COCO, and the excellent results demonstrate the promising performance of our proposed method and learning manner.
Authors: Jiyong Oh, Junhaeng Lee, Woongchan Byun, Minsang Kong, Sang Hun Lee
Abstract: Recent studies have focused on enhancing the performance of 3D object detection models. Among various approaches, ground-truth sampling has been proposed as an augmentation technique to address the challenges posed by limited ground-truth data. However, an inherent issue with ground-truth sampling is its tendency to increase false positives. Therefore, this study aims to overcome the limitations of ground-truth sampling and improve the performance of 3D object detection models by developing a new augmentation technique called false-positive sampling. False-positive sampling involves retraining the model using point clouds that are identified as false positives in the model's predictions. We propose an algorithm that utilizes both ground-truth and false-positive sampling and an algorithm for building the false-positive sample database. Additionally, we analyze the principles behind the performance enhancement due to false-positive sampling and propose a technique that applies the concept of curriculum learning to the sampling strategy that encompasses both false-positive and ground-truth sampling techniques. Our experiments demonstrate that models utilizing false-positive sampling show a reduction in false positives and exhibit improved object detection performance. On the KITTI and Waymo Open datasets, models with false-positive sampling surpass the baseline models by a large margin.
Authors: Cong Ma, Lei Qiao, Chengkai Zhu, Kai Liu, Zelong Kong, Qing Li, Xueqi Zhou, Yuheng Kan, Wei Wu
Abstract: Vehicle-to-everything (V2X) is a popular topic in the field of Autonomous Driving in recent years. Vehicle-infrastructure cooperation (VIC) becomes one of the important research area. Due to the complexity of traffic conditions such as blind spots and occlusion, it greatly limits the perception capabilities of single-view roadside sensing systems. To further enhance the accuracy of roadside perception and provide better information to the vehicle side, in this paper, we constructed holographic intersections with various layouts to build a large-scale multi-sensor holographic vehicle-infrastructure cooperation dataset, called HoloVIC. Our dataset includes 3 different types of sensors (Camera, Lidar, Fisheye) and employs 4 sensor-layouts based on the different intersections. Each intersection is equipped with 6-18 sensors to capture synchronous data. While autonomous vehicles pass through these intersections for collecting VIC data. HoloVIC contains in total on 100k+ synchronous frames from different sensors. Additionally, we annotated 3D bounding boxes based on Camera, Fisheye, and Lidar. We also associate the IDs of the same objects across different devices and consecutive frames in sequence. Based on HoloVIC, we formulated four tasks to facilitate the development of related research. We also provide benchmarks for these tasks.
Authors: Zhongqi Yue, Pan Zhou, Richang Hong, Hanwang Zhang, Qianru Sun
Abstract: Even when using large multi-modal foundation models, few-shot learning is still challenging -- if there is no proper inductive bias, it is nearly impossible to keep the nuanced class attributes while removing the visually prominent attributes that spuriously correlate with class labels. To this end, we find an inductive bias that the time-steps of a Diffusion Model (DM) can isolate the nuanced class attributes, i.e., as the forward diffusion adds noise to an image at each time-step, nuanced attributes are usually lost at an earlier time-step than the spurious attributes that are visually prominent. Building on this, we propose Time-step Few-shot (TiF) learner. We train class-specific low-rank adapters for a text-conditioned DM to make up for the lost attributes, such that images can be accurately reconstructed from their noisy ones given a prompt. Hence, at a small time-step, the adapter and prompt are essentially a parameterization of only the nuanced class attributes. For a test image, we can use the parameterization to only extract the nuanced class attributes for classification. TiF learner significantly outperforms OpenCLIP and its adapters on a variety of fine-grained and customized few-shot learning tasks. Codes are in https://github.com/yue-zhongqi/tif.
Authors: Weizhi Wang, Khalil Mrini, Linjie Yang, Sateesh Kumar, Yu Tian, Xifeng Yan, Heng Wang
Abstract: We propose a novel framework for filtering image-text data by leveraging fine-tuned Multimodal Language Models (MLMs). Our approach outperforms predominant filtering methods (e.g., CLIPScore) via integrating the recent advances in MLMs. We design four distinct yet complementary metrics to holistically measure the quality of image-text data. A new pipeline is established to construct high-quality instruction data for fine-tuning MLMs as data filters. Comparing with CLIPScore, our MLM filters produce more precise and comprehensive scores that directly improve the quality of filtered data and boost the performance of pre-trained models. We achieve significant improvements over CLIPScore on popular foundation models (i.e., CLIP and BLIP2) and various downstream tasks. Our MLM filter can generalize to different models and tasks, and be used as a drop-in replacement for CLIPScore. An additional ablation study is provided to verify our design choices for the MLM filter.
Authors: Yaoyan Zheng, Hongyu Yang, Di Huang
Abstract: Recent advancements in video semantic segmentation have made substantial progress by exploiting temporal correlations. Nevertheless, persistent challenges, including redundant computation and the reliability of the feature propagation process, underscore the need for further innovation. In response, we present Deep Common Feature Mining (DCFM), a novel approach strategically designed to address these challenges by leveraging the concept of feature sharing. DCFM explicitly decomposes features into two complementary components. The common representation extracted from a key-frame furnishes essential high-level information to neighboring non-key frames, allowing for direct re-utilization without feature propagation. Simultaneously, the independent feature, derived from each video frame, captures rapidly changing information, providing frame-specific clues crucial for segmentation. To achieve such decomposition, we employ a symmetric training strategy tailored for sparsely annotated data, empowering the backbone to learn a robust high-level representation enriched with common information. Additionally, we incorporate a self-supervised loss function to reinforce intra-class feature similarity and enhance temporal consistency. Experimental evaluations on the VSPW and Cityscapes datasets demonstrate the effectiveness of our method, showing a superior balance between accuracy and efficiency.
Authors: Gang Liu, Hongyang Li, Zerui He, Shenjun Zhong
Abstract: Leveraging pre-trained visual language models has become a widely adopted approach for improving performance in downstream visual question answering (VQA) applications. However, in the specialized field of medical VQA, the scarcity of available data poses a significant barrier to achieving reliable model generalization. Numerous methods have been proposed to enhance model generalization, addressing the issue from data-centric and model-centric perspectives. Data augmentation techniques are commonly employed to enrich the dataset, while various regularization approaches aim to prevent model overfitting, especially when training on limited data samples. In this paper, we introduce a method that incorporates gradient-guided parameter perturbations to the visual encoder of the multimodality model during both pre-training and fine-tuning phases, to improve model generalization for downstream medical VQA tasks. The small perturbation is adaptively generated by aligning with the direction of the moving average gradient in the optimization landscape, which is opposite to the directions of the optimizer's historical updates. It is subsequently injected into the model's visual encoder. The results show that, even with a significantly smaller pre-training image caption dataset, our approach achieves competitive outcomes on both VQA-RAD and SLAKE datasets.
Authors: Jiawei Hou, Xiaoyan Li, Wenhao Guan, Gang Zhang, Di Feng, Yuheng Du, Xiangyang Xue, Jian Pu
Abstract: In autonomous driving, 3D occupancy prediction outputs voxel-wise status and semantic labels for more comprehensive understandings of 3D scenes compared with traditional perception tasks, such as 3D object detection and bird's-eye view (BEV) semantic segmentation. Recent researchers have extensively explored various aspects of this task, including view transformation techniques, ground-truth label generation, and elaborate network design, aiming to achieve superior performance. However, the inference speed, crucial for running on an autonomous vehicle, is neglected. To this end, a new method, dubbed FastOcc, is proposed. By carefully analyzing the network effect and latency from four parts, including the input image resolution, image backbone, view transformation, and occupancy prediction head, it is found that the occupancy prediction head holds considerable potential for accelerating the model while keeping its accuracy. Targeted at improving this component, the time-consuming 3D convolution network is replaced with a novel residual-like architecture, where features are mainly digested by a lightweight 2D BEV convolution network and compensated by integrating the 3D voxel features interpolated from the original image features. Experiments on the Occ3D-nuScenes benchmark demonstrate that our FastOcc achieves state-of-the-art results with a fast inference speed.
Authors: Feng Hou, Jin Yuan, Ying Yang, Yang Liu, Yang Zhang, Cheng Zhong, Zhongchao Shi, Jianping Fan, Yong Rui, Zhiqiang He
Abstract: Traditional cross-domain tasks, including domain adaptation and domain generalization, rely heavily on training model by source domain data. With the recent advance of vision-language models (VLMs), viewed as natural source models, the cross-domain task changes to directly adapt the pre-trained source model to arbitrary target domains equipped with prior domain knowledge, and we name this task Adaptive Domain Generalization (ADG). However, current cross-domain datasets have many limitations, such as unrealistic domains, unclear domain definitions, and the inability to fine-grained domain decomposition, which drives us to establish a novel dataset DomainVerse for ADG. Benefiting from the introduced hierarchical definition of domain shifts, DomainVerse consists of about 0.5 million images from 390 fine-grained realistic domains. With the help of the constructed DomainVerse and VLMs, we propose two methods called Domain CLIP and Domain++ CLIP for tuning-free adaptive domain generalization. Extensive and comprehensive experiments demonstrate the significance of the dataset and the effectiveness of the proposed methods.
Authors: Akram Zaytar, Caleb Robinson, Gilles Q. Hacheme, Girmaw A. Tadesse, Rahul Dodhia, Juan M. Lavista Ferres, Lacey F. Hughey, Jared A. Stabach, Irene Amoke
Abstract: Rare object detection is a fundamental task in applied geospatial machine learning, however is often challenging due to large amounts of high-resolution satellite or aerial imagery and few or no labeled positive samples to start with. This paper addresses the problem of bootstrapping such a rare object detection task assuming there is no labeled data and no spatial prior over the area of interest. We propose novel offline and online cluster-based approaches for sampling patches that are significantly more efficient, in terms of exposing positive samples to a human annotator, than random sampling. We apply our methods for identifying bomas, or small enclosures for herd animals, in the Serengeti Mara region of Kenya and Tanzania. We demonstrate a significant enhancement in detection efficiency, achieving a positive sampling rate increase from 2% (random) to 30%. This advancement enables effective machine learning mapping even with minimal labeling budgets, exemplified by an F1 score on the boma detection task of 0.51 with a budget of 300 total patches.
Authors: Zhuohong Li, Wei He, Jiepan Li, Fangxiao Lu, Hongyan Zhang
Abstract: Large-scale high-resolution (HR) land-cover mapping is a vital task to survey the Earth's surface and resolve many challenges facing humanity. However, it is still a non-trivial task hindered by complex ground details, various landforms, and the scarcity of accurate training labels over a wide-span geographic area. In this paper, we propose an efficient, weakly supervised framework (Paraformer), a.k.a. Low-to-High Network (L2HNet) V2, to guide large-scale HR land-cover mapping with easy-access historical land-cover data of low resolution (LR). Specifically, existing land-cover mapping approaches reveal the dominance of CNNs in preserving local ground details but still suffer from insufficient global modeling in various landforms. Therefore, we design a parallel CNN-Transformer feature extractor in Paraformer, consisting of a downsampling-free CNN branch and a Transformer branch, to jointly capture local and global contextual information. Besides, facing the spatial mismatch of training data, a pseudo-label-assisted training (PLAT) module is adopted to reasonably refine LR labels for weakly supervised semantic segmentation of HR images. Experiments on two large-scale datasets demonstrate the superiority of Paraformer over other state-of-the-art methods for automatically updating HR land-cover maps from LR historical labels.
Authors: Chihiro Nakatani, Hiroaki Kawashima, Norimichi Ukita
Abstract: This paper proposes Group Activity Feature (GAF) learning in which features of multi-person activity are learned as a compact latent vector. Unlike prior work in which the manual annotation of group activities is required for supervised learning, our method learns the GAF through person attribute prediction without group activity annotations. By learning the whole network in an end-to-end manner so that the GAF is required for predicting the person attributes of people in a group, the GAF is trained as the features of multi-person activity. As a person attribute, we propose to use a person's action class and appearance features because the former is easy to annotate due to its simpleness, and the latter requires no manual annotation. In addition, we introduce a location-guided attribute prediction to disentangle the complex GAF for extracting the features of each target person properly. Various experimental results validate that our method outperforms SOTA methods quantitatively and qualitatively on two public datasets. Visualization of our GAF also demonstrates that our method learns the GAF representing fined-grained group activity classes. Code: https://github.com/chihina/GAFL-CVPR2024.
Authors: Cheng Huang, Shoudong Han, Mengyu He, Wenbo Zheng, Yuhao Wei
Abstract: Accurate data association is crucial in reducing confusion, such as ID switches and assignment errors, in multi-object tracking (MOT). However, existing advanced methods often overlook the diversity among trajectories and the ambiguity and conflicts present in motion and appearance cues, leading to confusion among detections, trajectories, and associations when performing simple global data association. To address this issue, we propose a simple, versatile, and highly interpretable data association approach called Decomposed Data Association (DDA). DDA decomposes the traditional association problem into multiple sub-problems using a series of non-learning-based modules and selectively addresses the confusion in each sub-problem by incorporating targeted exploitation of new cues. Additionally, we introduce Occlusion-aware Non-Maximum Suppression (ONMS) to retain more occluded detections, thereby increasing opportunities for association with trajectories and indirectly reducing the confusion caused by missed detections. Finally, based on DDA and ONMS, we design a powerful multi-object tracker named DeconfuseTrack, specifically focused on resolving confusion in MOT. Extensive experiments conducted on the MOT17 and MOT20 datasets demonstrate that our proposed DDA and ONMS significantly enhance the performance of several popular trackers. Moreover, DeconfuseTrack achieves state-of-the-art performance on the MOT17 and MOT20 test sets, significantly outperforms the baseline tracker ByteTrack in metrics such as HOTA, IDF1, AssA. This validates that our tracking design effectively reduces confusion caused by simple global association.
Authors: Yichen Yao, Zimo Jiang, Yujing Sun, Zhencai Zhu, Xinge Zhu, Runnan Chen, Yuexin Ma
Abstract: Human-centric 3D scene understanding has recently drawn increasing attention, driven by its critical impact on robotics. However, human-centric real-life scenarios are extremely diverse and complicated, and humans have intricate motions and interactions. With limited labeled data, supervised methods are difficult to generalize to general scenarios, hindering real-life applications. Mimicking human intelligence, we propose an unsupervised 3D detection method for human-centric scenarios by transferring the knowledge from synthetic human instances to real scenes. To bridge the gap between the distinct data representations and feature distributions of synthetic models and real point clouds, we introduce novel modules for effective instance-to-scene representation transfer and synthetic-to-real feature alignment. Remarkably, our method exhibits superior performance compared to current state-of-the-art techniques, achieving a substantial 87.8\% improvement in mAP and closely approaching the performance of fully supervised methods (62.15 mAP vs. 69.02 mAP) on HuCenLife.
Authors: Zheng Li, Xiang Li, Xinyi Fu, Xing Zhang, Weiqiang Wang, Jian Yang
Abstract: Prompt learning has emerged as a valuable technique in enhancing vision-language models (VLMs) such as CLIP for downstream tasks in specific domains. Existing work mainly focuses on designing various learning forms of prompts, neglecting the potential of prompts as effective distillers for learning from larger teacher models. In this paper, we introduce an unsupervised domain prompt distillation framework, which aims to transfer the knowledge of a larger teacher model to a lightweight target model through prompt-driven imitation using unlabeled domain images. Specifically, our framework consists of two distinct stages. In the initial stage, we pre-train a large CLIP teacher model using domain (few-shot) labels. After pre-training, we leverage the unique decoupled-modality characteristics of CLIP by pre-computing and storing the text features as class vectors only once through the teacher text encoder. In the subsequent stage, the stored class vectors are shared across teacher and student image encoders for calculating the predicted logits. Further, we align the logits of both the teacher and student models via KL divergence, encouraging the student image encoder to generate similar probability distributions to the teacher through the learnable prompts. The proposed prompt distillation process eliminates the reliance on labeled data, enabling the algorithm to leverage a vast amount of unlabeled images within the domain. Finally, the well-trained student image encoders and pre-stored text features (class vectors) are utilized for inference. To our best knowledge, we are the first to (1) perform unsupervised domain-specific prompt-driven knowledge distillation for CLIP, and (2) establish a practical pre-storing mechanism of text features as shared class vectors between teacher and student. Extensive experiments on 11 datasets demonstrate the effectiveness of our method.
Authors: Kumaranage Ravindu Yasas Nagasinghe, Honglu Zhou, Malitha Gunawardhana, Martin Renqiang Min, Daniel Harari, Muhammad Haris Khan
Abstract: In this paper, we explore the capability of an agent to construct a logical sequence of action steps, thereby assembling a strategic procedural plan. This plan is crucial for navigating from an initial visual observation to a target visual outcome, as depicted in real-life instructional videos. Existing works have attained partial success by extensively leveraging various sources of information available in the datasets, such as heavy intermediate visual observations, procedural names, or natural language step-by-step instructions, for features or supervision signals. However, the task remains formidable due to the implicit causal constraints in the sequencing of steps and the variability inherent in multiple feasible plans. To tackle these intricacies that previous efforts have overlooked, we propose to enhance the capabilities of the agent by infusing it with procedural knowledge. This knowledge, sourced from training procedure plans and structured as a directed weighted graph, equips the agent to better navigate the complexities of step sequencing and its potential variations. We coin our approach KEPP, a novel Knowledge-Enhanced Procedure Planning system, which harnesses a probabilistic procedural knowledge graph extracted from training data, effectively acting as a comprehensive textbook for the training domain. Experimental evaluations across three widely-used datasets under settings of varying complexity reveal that KEPP attains superior, state-of-the-art results while requiring only minimal supervision.
Authors: Lingyan Ran, Lushuang Wang, Tao Zhuo, Yinghui Xing
Abstract: Semantic segmentation of remote sensing images is a challenging and hot issue due to the large amount of unlabeled data. Unsupervised domain adaptation (UDA) has proven to be advantageous in incorporating unclassified information from the target domain. However, independently fine-tuning UDA models on the source and target domains has a limited effect on the outcome. This paper proposes a hybrid training strategy as well as a novel dual-domain image fusion strategy that effectively utilizes the original image, transformation image, and intermediate domain information. Moreover, to enhance the precision of pseudo-labels, we present a pseudo-label region-specific weight strategy. The efficacy of our approach is substantiated by extensive benchmark experiments and ablation studies conducted on the ISPRS Vaihingen and Potsdam datasets.
Authors: Yu Qiao, Apurba Adhikary, Chaoning Zhang, Choong Seon Hong
Abstract: Federated learning (FL) is a privacy-preserving distributed management framework based on collaborative model training of distributed devices in edge networks. However, recent studies have shown that FL is vulnerable to adversarial examples (AEs), leading to a significant drop in its performance. Meanwhile, the non-independent and identically distributed (non-IID) challenge of data distribution between edge devices can further degrade the performance of models. Consequently, both AEs and non-IID pose challenges to deploying robust learning models at the edge. In this work, we adopt the adversarial training (AT) framework to improve the robustness of FL models against adversarial example (AE) attacks, which can be termed as federated adversarial training (FAT). Moreover, we address the non-IID challenge by implementing a simple yet effective logits calibration strategy under the FAT framework, which can enhance the robustness of models when subjected to adversarial attacks. Specifically, we employ a direct strategy to adjust the logits output by assigning higher weights to classes with small samples during training. This approach effectively tackles the class imbalance in the training data, with the goal of mitigating biases between local and global models. Experimental results on three dataset benchmarks, MNIST, Fashion-MNIST, and CIFAR-10 show that our strategy achieves competitive results in natural and robust accuracy compared to several baselines.
Authors: Chenqiang Gao, Chuandong Liu, Jun Shu, Fangcen Liu, Jiang Liu, Luyu Yang, Xinbo Gao, Deyu Meng
Abstract: Current state-of-the-art (SOTA) 3D object detection methods often require a large amount of 3D bounding box annotations for training. However, collecting such large-scale densely-supervised datasets is notoriously costly. To reduce the cumbersome data annotation process, we propose a novel sparsely-annotated framework, in which we just annotate one 3D object per scene. Such a sparse annotation strategy could significantly reduce the heavy annotation burden, while inexact and incomplete sparse supervision may severely deteriorate the detection performance. To address this issue, we develop the SS3D++ method that alternatively improves 3D detector training and confident fully-annotated scene generation in a unified learning scheme. Using sparse annotations as seeds, we progressively generate confident fully-annotated scenes based on designing a missing-annotated instance mining module and reliable background mining module. Our proposed method produces competitive results when compared with SOTA weakly-supervised methods using the same or even more annotation costs. Besides, compared with SOTA fully-supervised methods, we achieve on-par or even better performance on the KITTI dataset with about 5x less annotation cost, and 90% of their performance on the Waymo dataset with about 15x less annotation cost. The additional unlabeled training scenes could further boost the performance. The code will be available at https://github.com/gaocq/SS3D2.
Authors: Weijie Li, Litong Gong, Yiran Zhu, Fanda Fan, Biao Wang, Tiezheng Ge, Bo Zheng
Abstract: Image-to-video (I2V) generation tasks always suffer from keeping high fidelity in the open domains. Traditional image animation techniques primarily focus on specific domains such as faces or human poses, making them difficult to generalize to open domains. Several recent I2V frameworks based on diffusion models can generate dynamic content for open domain images but fail to maintain fidelity. We found that two main factors of low fidelity are the loss of image details and the noise prediction biases during the denoising process. To this end, we propose an effective method that can be applied to mainstream video diffusion models. This method achieves high fidelity based on supplementing more precise image information and noise rectification. Specifically, given a specified image, our method first adds noise to the input image latent to keep more details, then denoises the noisy latent with proper rectification to alleviate the noise prediction biases. Our method is tuning-free and plug-and-play. The experimental results demonstrate the effectiveness of our approach in improving the fidelity of generated videos. For more image-to-video generated results, please refer to the project website: https://noise-rectification.github.io.
Authors: Philipp J. R\"osch, Norbert Oswald, Michaela Geierhos, Jind\v{r}ich Libovick\'y
Abstract: Current multimodal models leveraging contrastive learning often face limitations in developing fine-grained conceptual understanding. This is due to random negative samples during pretraining, causing almost exclusively very dissimilar concepts to be compared in the loss function. Consequently, the models struggle with fine-grained semantic differences. To address this problem, we introduce a novel pretraining method incorporating synthetic hard negative text examples. The hard negatives permute terms corresponding to visual concepts, leading to a more fine-grained visual and textual concept alignment. Further, we introduce InpaintCOCO, a new challenging dataset for assessing the fine-grained alignment of colors, objects, and sizes in vision-language models. We created the dataset using generative inpainting from COCO images by changing the visual concepts so that the images no longer match their original captions. Our results show significant improvements in fine-grained concept understanding across a wide range of vision-language datasets, including our InpaintCOCO dataset.
Authors: Han Lu, Xiaosong Jia, Yichen Xie, Wenlong Liao, Xiaokang Yang, Junchi Yan
Abstract: End-to-end differentiable learning for autonomous driving (AD) has recently become a prominent paradigm. One main bottleneck lies in its voracious appetite for high-quality labeled data e.g. 3D bounding boxes and semantic segmentation, which are notoriously expensive to manually annotate. The difficulty is further pronounced due to the prominent fact that the behaviors within samples in AD often suffer from long tailed distribution. In other words, a large part of collected data can be trivial (e.g. simply driving forward in a straight road) and only a few cases are safety-critical. In this paper, we explore a practically important yet under-explored problem about how to achieve sample and label efficiency for end-to-end AD. Specifically, we design a planning-oriented active learning method which progressively annotates part of collected raw data according to the proposed diversity and usefulness criteria for planning routes. Empirically, we show that our planning-oriented approach could outperform general active learning methods by a large margin. Notably, our method achieves comparable performance with state-of-the-art end-to-end AD methods - by using only 30% nuScenes data. We hope our work could inspire future works to explore end-to-end AD from a data-centric perspective in addition to methodology efforts.
Authors: Jinhong He, Minglong Xue, Zhipu Liu, Chengyun Song, Senming Zhong
Abstract: Diffusion model-based low-light image enhancement methods rely heavily on paired training data, leading to limited extensive application. Meanwhile, existing unsupervised methods lack effective bridging capabilities for unknown degradation. To address these limitations, we propose a novel zero-reference lighting estimation diffusion model for low-light image enhancement called Zero-LED. It utilizes the stable convergence ability of diffusion models to bridge the gap between low-light domains and real normal-light domains and successfully alleviates the dependence on pairwise training data via zero-reference learning. Specifically, we first design the initial optimization network to preprocess the input image and implement bidirectional constraints between the diffusion model and the initial optimization network through multiple objective functions. Subsequently, the degradation factors of the real-world scene are optimized iteratively to achieve effective light enhancement. In addition, we explore a frequency-domain based and semantically guided appearance reconstruction module that encourages feature alignment of the recovered image at a fine-grained level and satisfies subjective expectations. Finally, extensive experiments demonstrate the superiority of our approach to other state-of-the-art methods and more significant generalization capabilities. We will open the source code upon acceptance of the paper.
Authors: Fei Zhu, Xu-Yao Zhang, Zhen Cheng, Cheng-Lin Liu
Abstract: Reliable confidence estimation is a challenging yet fundamental requirement in many risk-sensitive applications. However, modern deep neural networks are often overconfident for their incorrect predictions, i.e., misclassified samples from known classes, and out-of-distribution (OOD) samples from unknown classes. In recent years, many confidence calibration and OOD detection methods have been developed. In this paper, we find a general, widely existing but actually-neglected phenomenon that most confidence estimation methods are harmful for detecting misclassification errors. We investigate this problem and reveal that popular calibration and OOD detection methods often lead to worse confidence separation between correctly classified and misclassified examples, making it difficult to decide whether to trust a prediction or not. Finally, we propose to enlarge the confidence gap by finding flat minima, which yields state-of-the-art failure prediction performance under various settings including balanced, long-tailed, and covariate-shift classification scenarios. Our study not only provides a strong baseline for reliable confidence estimation but also acts as a bridge between understanding calibration, OOD detection, and failure prediction. The code is available at \url{https://github.com/Impression2805/FMFP}.
Authors: Daniele Mari, Simone Milani
Abstract: Learned image compression codecs have recently achieved impressive compression performances surpassing the most efficient image coding architectures. However, most approaches are trained to minimize rate and distortion which often leads to unsatisfactory visual results at low bitrates since perceptual metrics are not taken into account. In this paper, we show that conditional diffusion models can lead to promising results in the generative compression task when used as a decoder, and that, given a compressed representation, they allow creating new tradeoff points between distortion and perception at the decoder side based on the sampling method.
Authors: Duy Tran Thanh, Yeejin Lee, Byeongkeun Kang
Abstract: This work addresses the task of long-term person re-identification. Typically, person re-identification assumes that people do not change their clothes, which limits its applications to short-term scenarios. To overcome this limitation, we investigate long-term person re-identification, which considers both clothes-changing and clothes-consistent scenarios. In this paper, we propose a novel framework that effectively learns and utilizes both global and local information. The proposed framework consists of three streams: global, local body part, and head streams. The global and head streams encode identity-relevant information from an entire image and a cropped image of the head region, respectively. Both streams encode the most distinct, less distinct, and average features using the combinations of adversarial erasing, max pooling, and average pooling. The local body part stream extracts identity-related information for each body part, allowing it to be compared with the same body part from another image. Since body part annotations are not available in re-identification datasets, pseudo-labels are generated using clustering. These labels are then utilized to train a body part segmentation head in the local body part stream. The proposed framework is trained by backpropagating the weighted summation of the identity classification loss, the pair-based loss, and the pseudo body part segmentation loss. To demonstrate the effectiveness of the proposed method, we conducted experiments on three publicly available datasets (Celeb-reID, PRCC, and VC-Clothes). The experimental results demonstrate that the proposed method outperforms the previous state-of-the-art method.
Authors: Abeer Banerjee, Naval K. Mehta, Shyam S. Prasad, Himanshu, Sumeet Saurav, Sanjay Singh
Abstract: In this paper, we address the intricate challenge of gaze vector prediction, a pivotal task with applications ranging from human-computer interaction to driver monitoring systems. Our innovative approach is designed for the demanding setting of extremely low-light conditions, leveraging a novel temporal event encoding scheme, and a dedicated neural network architecture. The temporal encoding method seamlessly integrates Dynamic Vision Sensor (DVS) events with grayscale guide frames, generating consecutively encoded images for input into our neural network. This unique solution not only captures diverse gaze responses from participants within the active age group but also introduces a curated dataset tailored for low-light conditions. The encoded temporal frames paired with our network showcase impressive spatial localization and reliable gaze direction in their predictions. Achieving a remarkable 100-pixel accuracy of 100%, our research underscores the potency of our neural network to work with temporally consecutive encoded images for precise gaze vector predictions in challenging low-light videos, contributing to the advancement of gaze prediction technologies.
Authors: Xijia Tao, Shuai Zhong, Lei Li, Qi Liu, Lingpeng Kong
Abstract: There has been an increasing interest in the alignment of large language models (LLMs) with human values. However, the safety issues of their integration with a vision module, or vision language models (VLMs), remain relatively underexplored. In this paper, we propose a novel jailbreaking attack against VLMs, aiming to bypass their safety barrier when a user inputs harmful instructions. A scenario where our poisoned (image, text) data pairs are included in the training data is assumed. By replacing the original textual captions with malicious jailbreak prompts, our method can perform jailbreak attacks with the poisoned images. Moreover, we analyze the effect of poison ratios and positions of trainable parameters on our attack's success rate. For evaluation, we design two metrics to quantify the success rate and the stealthiness of our attack. Together with a list of curated harmful instructions, a benchmark for measuring attack efficacy is provided. We demonstrate the efficacy of our attack by comparing it with baseline methods.
Authors: Sho Shimotsumagari, Shumpei Takezaki, Daichi Haraguchi, Seiichi Uchida
Abstract: The purpose of this paper is to enable the conversion between machine-printed character images (i.e., font images) and handwritten character images through machine learning. For this purpose, we propose a novel unpaired image-to-image domain conversion method, CycleDM, which incorporates the concept of CycleGAN into the diffusion model. Specifically, CycleDM has two internal conversion models that bridge the denoising processes of two image domains. These conversion models are efficiently trained without explicit correspondence between the domains. By applying machine-printed and handwritten character images to the two modalities, CycleDM realizes the conversion between them. Our experiments for evaluating the converted images quantitatively and qualitatively found that ours performs better than other comparable approaches.
Authors: Hagyeong Lee, Minkyu Kim, Jun-Hyuk Kim, Seungeon Kim, Dokwan Oh, Jaeho Lee
Abstract: Recent advances in text-guided image compression have shown great potential to enhance the perceptual quality of reconstructed images. These methods, however, tend to have significantly degraded pixel-wise fidelity, limiting their practicality. To fill this gap, we develop a new text-guided image compression algorithm that achieves both high perceptual and pixel-wise fidelity. In particular, we propose a compression framework that leverages text information mainly by text-adaptive encoding and training with joint image-text loss. By doing so, we avoid decoding based on text-guided generative models -- known for high generative diversity -- and effectively utilize the semantic information of text at a global level. Experimental results on various datasets show that our method can achieve high pixel-level and perceptual quality, with either human- or machine-generated captions. In particular, our method outperforms all baselines in terms of LPIPS, with some room for even more improvements when we use more carefully generated captions.
Authors: Ahmad Hassanpour, Yasamin Kowsari, Hatef Otroshi Shahreza, Bian Yang, Sebastien Marcel
Abstract: This paper explores the application of large language models (LLMs), like ChatGPT, for biometric tasks. We specifically examine the capabilities of ChatGPT in performing biometric-related tasks, with an emphasis on face recognition, gender detection, and age estimation. Since biometrics are considered as sensitive information, ChatGPT avoids answering direct prompts, and thus we crafted a prompting strategy to bypass its safeguard and evaluate the capabilities for biometrics tasks. Our study reveals that ChatGPT recognizes facial identities and differentiates between two facial images with considerable accuracy. Additionally, experimental results demonstrate remarkable performance in gender detection and reasonable accuracy for the age estimation tasks. Our findings shed light on the promising potentials in the application of LLMs and foundation models for biometrics.
Authors: Junwen He, Yifan Wang, Lijun Wang, Huchuan Lu, Jun-Yan He, Jin-Peng Lan, Bin Luo, Xuansong Xie
Abstract: Multimodal Large Language Model (MLLMs) leverages Large Language Models as a cognitive framework for diverse visual-language tasks. Recent efforts have been made to equip MLLMs with visual perceiving and grounding capabilities. However, there still remains a gap in providing fine-grained pixel-level perceptions and extending interactions beyond text-specific inputs. In this work, we propose {\bf{AnyRef}}, a general MLLM model that can generate pixel-wise object perceptions and natural language descriptions from multi-modality references, such as texts, boxes, images, or audio. This innovation empowers users with greater flexibility to engage with the model beyond textual and regional prompts, without modality-specific designs. Through our proposed refocusing mechanism, the generated grounding output is guided to better focus on the referenced object, implicitly incorporating additional pixel-level supervision. This simple modification utilizes attention scores generated during the inference of LLM, eliminating the need for extra computations while exhibiting performance enhancements in both grounding masks and referring expressions. With only publicly available training data, our model achieves state-of-the-art results across multiple benchmarks, including diverse modality referring segmentation and region-level referring expression generation.
Authors: Xue Song, Jiequan Cui, Hanwang Zhang, Jingjing Chen, Richang Hong, Yu-Gang Jiang
Abstract: We study text-based image editing (TBIE) of a single image by counterfactual inference because it is an elegant formulation to precisely address the requirement: the edited image should retain the fidelity of the original one. Through the lens of the formulation, we find that the crux of TBIE is that existing techniques hardly achieve a good trade-off between editability and fidelity, mainly due to the overfitting of the single-image fine-tuning. To this end, we propose a Doubly Abductive Counterfactual inference framework (DAC). We first parameterize an exogenous variable as a UNet LoRA, whose abduction can encode all the image details. Second, we abduct another exogenous variable parameterized by a text encoder LoRA, which recovers the lost editability caused by the overfitted first abduction. Thanks to the second abduction, which exclusively encodes the visual transition from post-edit to pre-edit, its inversion -- subtracting the LoRA -- effectively reverts pre-edit back to post-edit, thereby accomplishing the edit. Through extensive experiments, our DAC achieves a good trade-off between editability and fidelity. Thus, we can support a wide spectrum of user editing intents, including addition, removal, manipulation, replacement, style transfer, and facial change, which are extensively validated in both qualitative and quantitative evaluations. Codes are in https://github.com/xuesong39/DAC.
Authors: Jianjian Cao, Peng Ye, Shengze Li, Chong Yu, Yansong Tang, Jiwen Lu, Tao Chen
Abstract: Vision-Language Transformers (VLTs) have shown great success recently, but are meanwhile accompanied by heavy computation costs, where a major reason can be attributed to the large number of visual and language tokens. Existing token pruning research for compressing VLTs mainly follows a single-modality-based scheme yet ignores the critical role of aligning different modalities for guiding the token pruning process, causing the important tokens for one modality to be falsely pruned in another modality branch. Meanwhile, existing VLT pruning works also lack the flexibility to dynamically compress each layer based on different input samples. To this end, we propose a novel framework named Multimodal Alignment-Guided Dynamic Token Pruning (MADTP) for accelerating various VLTs. Specifically, we first introduce a well-designed Multi-modality Alignment Guidance (MAG) module that can align features of the same semantic concept from different modalities, to ensure the pruned tokens are less important for all modalities. We further design a novel Dynamic Token Pruning (DTP) module, which can adaptively adjust the token compression ratio in each layer based on different input instances. Extensive experiments on various benchmarks demonstrate that MADTP significantly reduces the computational complexity of kinds of multimodal models while preserving competitive performance. Notably, when applied to the BLIP model in the NLVR2 dataset, MADTP can reduce the GFLOPs by 80% with less than 4% performance degradation.
Authors: Yuheng Jia, Jianhong Cheng, Hui Liu, Junhui Hou
Abstract: Deep clustering has exhibited remarkable performance; however, the overconfidence problem, i.e., the estimated confidence for a sample belonging to a particular cluster greatly exceeds its actual prediction accuracy, has been overlooked in prior research. To tackle this critical issue, we pioneer the development of a calibrated deep clustering framework. Specifically, we propose a novel dual-head deep clustering pipeline that can effectively calibrate the estimated confidence and the actual accuracy. The calibration head adjusts the overconfident predictions of the clustering head using regularization methods, generating prediction confidence and pseudo-labels that match the model learning status. This calibration process also guides the clustering head in dynamically selecting reliable high-confidence samples for training. Additionally, we introduce an effective network initialization strategy that enhances both training speed and network robustness. Extensive experiments demonstrate the proposed calibrated deep clustering framework not only surpasses state-of-the-art deep clustering methods by approximately 10 times in terms of expected calibration error but also significantly outperforms them in terms of clustering accuracy.
Authors: Gen Luo, Yiyi Zhou, Yuxin Zhang, Xiawu Zheng, Xiaoshuai Sun, Rongrong Ji
Abstract: Despite remarkable progress, existing multimodal large language models (MLLMs) are still inferior in granular visual recognition. Contrary to previous works, we study this problem from the perspective of image resolution, and reveal that a combination of low- and high-resolution visual features can effectively mitigate this shortcoming. Based on this observation, we propose a novel and efficient method for MLLMs, termed Mixture-of-Resolution Adaptation (MRA). In particular, MRA adopts two visual pathways for images with different resolutions, where high-resolution visual information is embedded into the low-resolution pathway via the novel mixture-of-resolution adapters (MR-Adapters). This design also greatly reduces the input sequence length of MLLMs. To validate MRA, we apply it to a recent MLLM called LLaVA, and term the new model LLaVA-HR. We conduct extensive experiments on 11 vision-language (VL) tasks, which show that LLaVA-HR outperforms existing MLLMs on 8 VL tasks, e.g., +9.4% on TextVQA. More importantly, both training and inference of LLaVA-HR remain efficient with MRA, e.g., 20 training hours and 3$\times$ inference speed than LLaVA-1.5. Source codes are released at: https://github.com/luogen1996/LLaVA-HR.
Authors: Simone Alberto Peirone, Francesca Pistilli, Antonio Alliegro, Giuseppe Averta
Abstract: Human comprehension of a video stream is naturally broad: in a few instants, we are able to understand what is happening, the relevance and relationship of objects, and forecast what will follow in the near future, everything all at once. We believe that - to effectively transfer such an holistic perception to intelligent machines - an important role is played by learning to correlate concepts and to abstract knowledge coming from different tasks, to synergistically exploit them when learning novel skills. To accomplish this, we seek for a unified approach to video understanding which combines shared temporal modelling of human actions with minimal overhead, to support multiple downstream tasks and enable cooperation when learning novel skills. We then propose EgoPack, a solution that creates a collection of task perspectives that can be carried across downstream tasks and used as a potential source of additional insights, as a backpack of skills that a robot can carry around and use when needed. We demonstrate the effectiveness and efficiency of our approach on four Ego4D benchmarks, outperforming current state-of-the-art methods.
Authors: Zhen Yao, Jiawei Xu, Shuhang Hou, Mooi Choo Chuah
Abstract: Routine visual inspections of concrete structures are imperative for upholding the safety and integrity of critical infrastructure. Such visual inspections sometimes happen under low-light conditions, e.g., checking for bridge health. Crack segmentation under such conditions is challenging due to the poor contrast between cracks and their surroundings. However, most deep learning methods are designed for well-illuminated crack images and hence their performance drops dramatically in low-light scenes. In addition, conventional approaches require many annotated low-light crack images which is time-consuming. In this paper, we address these challenges by proposing CrackNex, a framework that utilizes reflectance information based on Retinex Theory to help the model learn a unified illumination-invariant representation. Furthermore, we utilize few-shot segmentation to solve the inefficient training data problem. In CrackNex, both a support prototype and a reflectance prototype are extracted from the support set. Then, a prototype fusion module is designed to integrate the features from both prototypes. CrackNex outperforms the SOTA methods on multiple datasets. Additionally, we present the first benchmark dataset, LCSD, for low-light crack segmentation. LCSD consists of 102 well-illuminated crack images and 41 low-light crack images. The dataset and code are available at https://github.com/zy1296/CrackNex.
Authors: Chun-Peng Chang, Shaoxiang Wang, Alain Pagani, Didier Stricker
Abstract: 3D visual grounding involves matching natural language descriptions with their corresponding objects in 3D spaces. Existing methods often face challenges with accuracy in object recognition and struggle in interpreting complex linguistic queries, particularly with descriptions that involve multiple anchors or are view-dependent. In response, we present the MiKASA (Multi-Key-Anchor Scene-Aware) Transformer. Our novel end-to-end trained model integrates a self-attention-based scene-aware object encoder and an original multi-key-anchor technique, enhancing object recognition accuracy and the understanding of spatial relationships. Furthermore, MiKASA improves the explainability of decision-making, facilitating error diagnosis. Our model achieves the highest overall accuracy in the Referit3D challenge for both the Sr3D and Nr3D datasets, particularly excelling by a large margin in categories that require viewpoint-dependent descriptions. The source code and additional resources for this project are available on GitHub: https://github.com/birdy666/MiKASA-3DVG
Authors: Yuxin Guo, Shijie Ma, Yuhao Zhao, Hu Su, Wei Zou
Abstract: Audio-Visual Source Localization (AVSL) is the task of identifying specific sounding objects in the scene given audio cues. In our work, we focus on semi-supervised AVSL with pseudo-labeling. To address the issues with vanilla hard pseudo-labels including bias accumulation, noise sensitivity, and instability, we propose a novel method named Cross Pseudo-Labeling (XPL), wherein two models learn from each other with the cross-refine mechanism to avoid bias accumulation. We equip XPL with two effective components. Firstly, the soft pseudo-labels with sharpening and pseudo-label exponential moving average mechanisms enable models to achieve gradual self-improvement and ensure stable training. Secondly, the curriculum data selection module adaptively selects pseudo-labels with high quality during training to mitigate potential bias. Experimental results demonstrate that XPL significantly outperforms existing methods, achieving state-of-the-art performance while effectively mitigating confirmation bias and ensuring training stability.
Authors: Mohamed Afifi, Mohamed ElHelw
Abstract: Perception is a key element for enabling intelligent autonomous navigation. Understanding the semantics of the surrounding environment and accurate vehicle pose estimation are essential capabilities for autonomous vehicles, including self-driving cars and mobile robots that perform complex tasks. Fast moving platforms like self-driving cars impose a hard challenge for localization and mapping algorithms. In this work, we propose a novel framework for real-time LiDAR odometry and mapping based on LOAM architecture for fast moving platforms. Our framework utilizes semantic information produced by a deep learning model to improve point-to-line and point-to-plane matching between LiDAR scans and build a semantic map of the environment, leading to more accurate motion estimation using LiDAR data. We observe that including semantic information in the matching process introduces a new type of outlier matches to the process, where matching occur between different objects of the same semantic class. To this end, we propose a novel algorithm that explicitly identifies and discards potential outliers in the matching process. In our experiments, we study the effect of improving the matching process on the robustness of LiDAR odometry against high speed motion. Our experimental evaluations on KITTI dataset demonstrate that utilizing semantic information and rejecting outliers significantly enhance the robustness of LiDAR odometry and mapping when there are large gaps between scan acquisition poses, which is typical for fast moving platforms.
Authors: Robert MendelRegensburg Medical Image Computing, Tobias RueckertRegensburg Medical Image Computing, Dirk WilhelmDepartment of Surgery, Faculty of Medicine, Klinikum rechts der Isar, Technical University of Munich, Munich, Germany, Daniel RueckertArtificial Intelligence in Healthcare and Medicine, Klinikum rechts der Isar, Technical University of Munich, Munich, Germany, Christoph PalmRegensburg Medical Image Computing
Abstract: Real-time computational speed and a high degree of precision are requirements for computer-assisted interventions. Applying a segmentation network to a medical video processing task can introduce significant inter-frame prediction noise. Existing approaches can reduce inconsistencies by including temporal information but often impose requirements on the architecture or dataset. This paper proposes a method to include temporal information in any segmentation model and, thus, a technique to improve video segmentation performance without alterations during training or additional labeling. With Motion-Corrected Moving Average, we refine the exponential moving average between the current and previous predictions. Using optical flow to estimate the movement between consecutive frames, we can shift the prior term in the moving-average calculation to align with the geometry of the current frame. The optical flow calculation does not require the output of the model and can therefore be performed in parallel, leading to no significant runtime penalty for our approach. We evaluate our approach on two publicly available segmentation datasets and two proprietary endoscopic datasets and show improvements over a baseline approach.
Authors: Yannan He, Garvita Tiwari, Tolga Birdal, Jan Eric Lenssen, Gerard Pons-Moll
Abstract: Faithfully modeling the space of articulations is a crucial task that allows recovery and generation of realistic poses, and remains a notorious challenge. To this end, we introduce Neural Riemannian Distance Fields (NRDFs), data-driven priors modeling the space of plausible articulations, represented as the zero-level-set of a neural field in a high-dimensional product-quaternion space. To train NRDFs only on positive examples, we introduce a new sampling algorithm, ensuring that the geodesic distances follow a desired distribution, yielding a principled distance field learning paradigm. We then devise a projection algorithm to map any random pose onto the level-set by an adaptive-step Riemannian optimizer, adhering to the product manifold of joint rotations at all times. NRDFs can compute the Riemannian gradient via backpropagation and by mathematical analogy, are related to Riemannian flow matching, a recent generative model. We conduct a comprehensive evaluation of NRDF against other pose priors in various downstream tasks, i.e., pose generation, image-based pose estimation, and solving inverse kinematics, highlighting NRDF's superior performance. Besides humans, NRDF's versatility extends to hand and animal poses, as it can effectively represent any articulation.
Authors: Kevin Shen, Surabhi S Nath, Aenne Brielmann, Peter Dayan
Abstract: The complexity of visual stimuli plays an important role in many cognitive phenomena, including attention, engagement, memorability, time perception and aesthetic evaluation. Despite its importance, complexity is poorly understood and ironically, previous models of image complexity have been quite \textit{complex}. There have been many attempts to find handcrafted features that explain complexity, but these features are usually dataset specific, and hence fail to generalise. On the other hand, more recent work has employed deep neural networks to predict complexity, but these models remain difficult to interpret, and do not guide a theoretical understanding of the problem. Here we propose to model complexity using segment-based representations of images. We use state-of-the-art segmentation models, SAM and FC-CLIP, to quantify the number of segments at multiple granularities, and the number of classes in an image respectively. We find that complexity is well-explained by a simple linear model with these two features across six diverse image-sets of naturalistic scene and art images. This suggests that the complexity of images can be surprisingly simple.
Authors: Yuxin Guo, Shijie Ma, Hu Su, Zhiqing Wang, Yuhao Zhao, Wei Zou, Siyang Sun, Yun Zheng
Abstract: Audio-Visual Source Localization (AVSL) aims to locate sounding objects within video frames given the paired audio clips. Existing methods predominantly rely on self-supervised contrastive learning of audio-visual correspondence. Without any bounding-box annotations, they struggle to achieve precise localization, especially for small objects, and suffer from blurry boundaries and false positives. Moreover, the naive semi-supervised method is poor in fully leveraging the information of abundant unlabeled data. In this paper, we propose a novel semi-supervised learning framework for AVSL, namely Dual Mean-Teacher (DMT), comprising two teacher-student structures to circumvent the confirmation bias issue. Specifically, two teachers, pre-trained on limited labeled data, are employed to filter out noisy samples via the consensus between their predictions, and then generate high-quality pseudo-labels by intersecting their confidence maps. The sufficient utilization of both labeled and unlabeled data and the proposed unbiased framework enable DMT to outperform current state-of-the-art methods by a large margin, with CIoU of 90.4% and 48.8% on Flickr-SoundNet and VGG-Sound Source, obtaining 8.9%, 9.6% and 4.6%, 6.4% improvements over self- and semi-supervised methods respectively, given only 3% positional-annotations. We also extend our framework to some existing AVSL methods and consistently boost their performance.
Authors: Kangning Cui, Zishan Shao, Gregory Larsen, Victor Pauca, Sarra Alqahtani, David Segurado, Jo\~ao Pinheiro, Manqi Wang, David Lutz, Robert Plemmons, Miles Silman
Abstract: Palms play an outsized role in tropical forests and are important resources for humans and wildlife. A central question in tropical ecosystems is understanding palm distribution and abundance. However, accurately identifying and localizing palms in geospatial imagery presents significant challenges due to dense vegetation, overlapping canopies, and variable lighting conditions in mixed-forest landscapes. Addressing this, we introduce PalmProbNet, a probabilistic approach utilizing transfer learning to analyze high-resolution UAV-derived orthomosaic imagery, enabling the detection of palm trees within the dense canopy of the Ecuadorian Rainforest. This approach represents a substantial advancement in automated palm detection, effectively pinpointing palm presence and locality in mixed tropical rainforests. Our process begins by generating an orthomosaic image from UAV images, from which we extract and label palm and non-palm image patches in two distinct sizes. These patches are then used to train models with an identical architecture, consisting of an unaltered pre-trained ResNet-18 and a Multilayer Perceptron (MLP) with specifically trained parameters. Subsequently, PalmProbNet employs a sliding window technique on the landscape orthomosaic, using both small and large window sizes to generate a probability heatmap. This heatmap effectively visualizes the distribution of palms, showcasing the scalability and adaptability of our approach in various forest densities. Despite the challenging terrain, our method demonstrated remarkable performance, achieving an accuracy of 97.32% and a Cohen's kappa of 94.59% in testing.
Authors: Ruizhuo Song, Beiming Yuan
Abstract: Abstract reasoning problems challenge the perceptual and cognitive abilities of AI algorithms, demanding deeper pattern discernment and inductive reasoning beyond explicit image features. This study introduces PMoC, a tailored probability model for the Bongard-Logo problem, achieving high reasoning accuracy by constructing independent probability models. Additionally, we present Pose-Transformer, an enhanced Transformer-Encoder designed for complex abstract reasoning tasks, including Bongard-Logo, RAVEN, I-RAVEN, and PGM. Pose-Transformer incorporates positional information learning, inspired by capsule networks' pose matrices, enhancing its focus on local positional relationships in image data processing. When integrated with PMoC, it further improves reasoning accuracy. Our approach effectively addresses reasoning difficulties associated with abstract entities' positional changes, outperforming previous models on the OIG, D3$\times$3 subsets of RAVEN, and PGM databases. This research contributes to advancing AI's capabilities in abstract reasoning and cognitive pattern recognition.
Authors: Ruizhuo Song, Beiming Yuan
Abstract: Abstract reasoning problems pose significant challenges to artificial intelligence algorithms, demanding cognitive capabilities beyond those required for perception tasks. This study introduces the Triple-CFN approach to tackle the Bongard-Logo problem, achieving notable reasoning accuracy by implicitly reorganizing the concept space of conflicting instances. Additionally, the Triple-CFN paradigm proves effective for the RPM problem with necessary modifications, yielding competitive results. To further enhance performance on the RPM issue, we develop the Meta Triple-CFN network, which explicitly structures the problem space while maintaining interpretability on progressive patterns. The success of Meta Triple-CFN is attributed to its paradigm of modeling the conceptual space, equivalent to normalizing reasoning information. Based on this ideology, we introduce the Re-space layer, enhancing the performance of both Meta Triple-CFN and Triple-CFN. This paper aims to contribute to advancements in machine intelligence by exploring innovative network designs for addressing abstract reasoning problems, paving the way for further breakthroughs in this domain.
Authors: Patrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari, Jonas M\"uller, Harry Saini, Yam Levi, Dominik Lorenz, Axel Sauer, Frederic Boesel, Dustin Podell, Tim Dockhorn, Zion English, Kyle Lacey, Alex Goodwin, Yannik Marek, Robin Rombach
Abstract: Diffusion models create data from noise by inverting the forward paths of data towards noise and have emerged as a powerful generative modeling technique for high-dimensional, perceptual data such as images and videos. Rectified flow is a recent generative model formulation that connects data and noise in a straight line. Despite its better theoretical properties and conceptual simplicity, it is not yet decisively established as standard practice. In this work, we improve existing noise sampling techniques for training rectified flow models by biasing them towards perceptually relevant scales. Through a large-scale study, we demonstrate the superior performance of this approach compared to established diffusion formulations for high-resolution text-to-image synthesis. Additionally, we present a novel transformer-based architecture for text-to-image generation that uses separate weights for the two modalities and enables a bidirectional flow of information between image and text tokens, improving text comprehension, typography, and human preference ratings. We demonstrate that this architecture follows predictable scaling trends and correlates lower validation loss to improved text-to-image synthesis as measured by various metrics and human evaluations. Our largest models outperform state-of-the-art models, and we will make our experimental data, code, and model weights publicly available.
Authors: Meng Zheng, Benjamin Planche, Xuan Gong, Fan Yang, Terrence Chen, Ziyan Wu
Abstract: 3D patient body modeling is critical to the success of automated patient positioning for smart medical scanning and operating rooms. Existing CNN-based end-to-end patient modeling solutions typically require a) customized network designs demanding large amount of relevant training data, covering extensive realistic clinical scenarios (e.g., patient covered by sheets), which leads to suboptimal generalizability in practical deployment, b) expensive 3D human model annotations, i.e., requiring huge amount of manual effort, resulting in systems that scale poorly. To address these issues, we propose a generic modularized 3D patient modeling method consists of (a) a multi-modal keypoint detection module with attentive fusion for 2D patient joint localization, to learn complementary cross-modality patient body information, leading to improved keypoint localization robustness and generalizability in a wide variety of imaging (e.g., CT, MRI etc.) and clinical scenarios (e.g., heavy occlusions); and (b) a self-supervised 3D mesh regression module which does not require expensive 3D mesh parameter annotations to train, bringing immediate cost benefits for clinical deployment. We demonstrate the efficacy of the proposed method by extensive patient positioning experiments on both public and clinical data. Our evaluation results achieve superior patient positioning performance across various imaging modalities in real clinical scenarios.
Authors: Chris Rockwell, Nilesh Kulkarni, Linyi Jin, Jeong Joon Park, Justin Johnson, David F. Fouhey
Abstract: Estimating relative camera poses between images has been a central problem in computer vision. Methods that find correspondences and solve for the fundamental matrix offer high precision in most cases. Conversely, methods predicting pose directly using neural networks are more robust to limited overlap and can infer absolute translation scale, but at the expense of reduced precision. We show how to combine the best of both methods; our approach yields results that are both precise and robust, while also accurately inferring translation scales. At the heart of our model lies a Transformer that (1) learns to balance between solved and learned pose estimations, and (2) provides a prior to guide a solver. A comprehensive analysis supports our design choices and demonstrates that our method adapts flexibly to various feature extractors and correspondence estimators, showing state-of-the-art performance in 6DoF pose estimation on Matterport3D, InteriorNet, StreetLearn, and Map-free Relocalization.
Authors: Ruirui Lin, Nantheera Anantrasirichai, Alexandra Malyugina, David Bull
Abstract: Distortions caused by low-light conditions are not only visually unpleasant but also degrade the performance of computer vision tasks. The restoration and enhancement have proven to be highly beneficial. However, there are only a limited number of enhancement methods explicitly designed for videos acquired in low-light conditions. We propose a Spatio-Temporal Aligned SUNet (STA-SUNet) model using a Swin Transformer as a backbone to capture low light video features and exploit their spatio-temporal correlations. The STA-SUNet model is trained on a novel, fully registered dataset (BVI), which comprises dynamic scenes captured under varying light conditions. It is further analysed comparatively against various other models over three test datasets. The model demonstrates superior adaptivity across all datasets, obtaining the highest PSNR and SSIM values. It is particularly effective in extreme low-light conditions, yielding fairly good visualisation results.
Authors: Yash Akhauri, Mohamed S. Abdelfattah
Abstract: Efficient deployment of neural networks (NN) requires the co-optimization of accuracy and latency. For example, hardware-aware neural architecture search has been used to automatically find NN architectures that satisfy a latency constraint on a specific hardware device. Central to these search algorithms is a prediction model that is designed to provide a hardware latency estimate for a candidate NN architecture. Recent research has shown that the sample efficiency of these predictive models can be greatly improved through pre-training on some \textit{training} devices with many samples, and then transferring the predictor on the \textit{test} (target) device. Transfer learning and meta-learning methods have been used for this, but often exhibit significant performance variability. Additionally, the evaluation of existing latency predictors has been largely done on hand-crafted training/test device sets, making it difficult to ascertain design features that compose a robust and general latency predictor. To address these issues, we introduce a comprehensive suite of latency prediction tasks obtained in a principled way through automated partitioning of hardware device sets. We then design a general latency predictor to comprehensively study (1) the predictor architecture, (2) NN sample selection methods, (3) hardware device representations, and (4) NN operation encoding schemes. Building on conclusions from our study, we present an end-to-end latency predictor training strategy that outperforms existing methods on 11 out of 12 difficult latency prediction tasks, improving latency prediction by 22.5\% on average, and up to to 87.6\% on the hardest tasks. Focusing on latency prediction, our HW-Aware NAS reports a $5.8\times$ speedup in wall-clock time. Our code is available on \href{https://github.com/abdelfattah-lab/nasflat_latency}{https://github.com/abdelfattah-lab/nasflat\_latency}.
URLs: https://github.com/abdelfattah-lab/nasflat_latency, https://github.com/abdelfattah-lab/nasflat\_latency
Authors: Yash Akhauri, Mohamed S. Abdelfattah
Abstract: Predictor-based methods have substantially enhanced Neural Architecture Search (NAS) optimization. The efficacy of these predictors is largely influenced by the method of encoding neural network architectures. While traditional encodings used an adjacency matrix describing the graph structure of a neural network, novel encodings embrace a variety of approaches from unsupervised pretraining of latent representations to vectors of zero-cost proxies. In this paper, we categorize and investigate neural encodings from three main types: structural, learned, and score-based. Furthermore, we extend these encodings and introduce \textit{unified encodings}, that extend NAS predictors to multiple search spaces. Our analysis draws from experiments conducted on over 1.5 million neural network architectures on NAS spaces such as NASBench-101 (NB101), NB201, NB301, Network Design Spaces (NDS), and TransNASBench-101. Building on our study, we present our predictor \textbf{FLAN}: \textbf{Fl}ow \textbf{A}ttention for \textbf{N}AS. FLAN integrates critical insights on predictor design, transfer learning, and \textit{unified encodings} to enable more than an order of magnitude cost reduction for training NAS accuracy predictors. Our implementation and encodings for all neural networks are open-sourced at \href{https://github.com/abdelfattah-lab/flan_nas}{https://github.com/abdelfattah-lab/flan\_nas}.
URLs: https://github.com/abdelfattah-lab/flan_nas, https://github.com/abdelfattah-lab/flan\_nas
Authors: Bardia Yousefi, M\'elina Khansari, Ryan Trask, Patrick Tallon, Carina Carino, Arman Afrasiyabi, Vikas Kundra, Lan Ma, Lei Ren, Keyvan Farahani, Michelle Hershman
Abstract: The isometric mapping method employs the shortest path algorithm to estimate the Euclidean distance between points on High dimensional (HD) manifolds. This may not be sufficient for weakly uniformed HD data as it could lead to overestimating distances between far neighboring points, resulting in inconsistencies between the intrinsic (local) and extrinsic (global) distances during the projection. To address this issue, we modify the shortest path algorithm by adding a novel constraint inspired by the Parzen-Rosenblatt (PR) window, which helps to maintain the uniformity of the constructed shortest-path graph in Isomap. Multiple imaging datasets overall of 72,236 cases, 70,000 MINST data, 1596 from multiple Chest-XRay pneumonia datasets, and three NSCLC CT/PET datasets with a total of 640 lung cancer patients, were used to benchmark and validate PR-Isomap. 431 imaging biomarkers were extracted from each modality. Our results indicate that PR-Isomap projects HD attributes into a lower-dimensional (LD) space while preserving information, visualized by the MNIST dataset indicating the maintaining local and global distances. PR-Isomap achieved the highest comparative accuracies of 80.9% (STD:5.8) for pneumonia and 78.5% (STD:4.4), 88.4% (STD:1.4), and 61.4% (STD:11.4) for three NSCLC datasets, with a confidence interval of 95% for outcome prediction. Similarly, the multivariate Cox model showed higher overall survival, measured with c-statistics and log-likelihood test, of PR-Isomap compared to other dimensionality reduction methods. Kaplan Meier survival curve also signifies the notable ability of PR-Isomap to distinguish between high-risk and low-risk patients using multimodal imaging biomarkers preserving HD imaging characteristics for precision medicine.
Authors: Mariusz Bujny, Katarzyna Jesionek, Jakub Nalepa, Karol Miszalski-Jamka, Katarzyna Widawka-\.Zak, Sabina Wolny, Marcin Kostur
Abstract: Precise localization of coronary arteries in Computed Tomography (CT) scans is critical from the perspective of medical assessment of coronary artery disease. Although various methods exist that offer high-quality segmentation of coronary arteries in cardiac contrast-enhanced CT scans, the potential of less invasive, non-contrast CT in this area is still not fully exploited. Since such fine anatomical structures are hardly visible in this type of medical images, the existing methods are characterized by high recall and low precision, and are used mainly for filtering of atherosclerotic plaques in the context of calcium scoring. In this paper, we address this research gap and introduce a deep learning algorithm for segmenting coronary arteries in multi-vendor ECG-gated non-contrast cardiac CT images which benefits from a novel framework for semi-automatic generation of Ground Truth (GT) via image registration. We hypothesize that the proposed GT generation process is much more efficient in this case than manual segmentation, since it allows for a fast generation of large volumes of diverse data, which leads to well-generalizing models. To investigate and thoroughly evaluate the segmentation quality based on such an approach, we propose a novel method for manual mesh-to-image registration, which is used to create our test-GT. The experimental study shows that the trained model has significantly higher accuracy than the GT used for training, and leads to the Dice and clDice metrics close to the interrater variability.
Authors: Brenda Y. Miao, Irene Y. Chen, Christopher YK Williams, Jays\'on Davidson, Augusto Garcia-Agundez, Harry Sun, Travis Zack, Atul J. Butte, Madhumita Sushil
Abstract: Recent advances in generative models, including large language models (LLMs), vision language models (VLMs), and diffusion models, have accelerated the field of natural language and image processing in medicine and marked a significant paradigm shift in how biomedical models can be developed and deployed. While these models are highly adaptable to new tasks, scaling and evaluating their usage presents new challenges not addressed in previous frameworks. In particular, the ability of these models to produce useful outputs with little to no specialized training data ("zero-" or "few-shot" approaches), as well as the open-ended nature of their outputs, necessitate the development of updated guidelines in using and evaluating these models. In response to gaps in standards and best practices for the development of clinical AI tools identified by US Executive Order 141103 and several emerging national networks for clinical AI evaluation, we begin to formalize some of these guidelines by building on the "Minimum information about clinical artificial intelligence modeling" (MI-CLAIM) checklist. The MI-CLAIM checklist, originally developed in 2020, provided a set of six steps with guidelines on the minimum information necessary to encourage transparent, reproducible research for artificial intelligence (AI) in medicine. Here, we propose modifications to the original checklist that highlight differences in training, evaluation, interpretability, and reproducibility of generative models compared to traditional AI models for clinical research. This updated checklist also seeks to clarify cohort selection reporting and adds additional items on alignment with ethical standards.
Authors: Zhaoxin Fan, Runmin Jiang, Junhao Wu, Xin Huang, Tianyang Wang, Heng Huang, Min Xu
Abstract: 3D medical image segmentation is a challenging task with crucial implications for disease diagnosis and treatment planning. Recent advances in deep learning have significantly enhanced fully supervised medical image segmentation. However, this approach heavily relies on labor-intensive and time-consuming fully annotated ground-truth labels, particularly for 3D volumes. To overcome this limitation, we propose a novel probabilistic-aware weakly supervised learning pipeline, specifically designed for 3D medical imaging. Our pipeline integrates three innovative components: a probability-based pseudo-label generation technique for synthesizing dense segmentation masks from sparse annotations, a Probabilistic Multi-head Self-Attention network for robust feature extraction within our Probabilistic Transformer Network, and a Probability-informed Segmentation Loss Function to enhance training with annotation confidence. Demonstrating significant advances, our approach not only rivals the performance of fully supervised methods but also surpasses existing weakly supervised methods in CT and MRI datasets, achieving up to 18.1% improvement in Dice scores for certain organs. The code is available at https://github.com/runminjiang/PW4MedSeg.
Authors: Zihao Wang, Rui Zhu, Dongruo Zhou, Zhikun Zhang, John Mitchell, Haixu Tang, XiaoFeng Wang
Abstract: Recent developments have underscored the critical role of \textit{differential privacy} (DP) in safeguarding individual data for training machine learning models. However, integrating DP oftentimes incurs significant model performance degradation due to the perturbation introduced into the training process, presenting a formidable challenge in the {differentially private machine learning} (DPML) field. To this end, several mitigative efforts have been proposed, typically revolving around formulating new DPML algorithms or relaxing DP definitions to harmonize with distinct contexts. In spite of these initiatives, the diminishment induced by DP on models, particularly large-scale models, remains substantial and thus, necessitates an innovative solution that adeptly circumnavigates the consequential impairment of model utility. In response, we introduce DPAdapter, a pioneering technique designed to amplify the model performance of DPML algorithms by enhancing parameter robustness. The fundamental intuition behind this strategy is that models with robust parameters are inherently more resistant to the noise introduced by DP, thereby retaining better performance despite the perturbations. DPAdapter modifies and enhances the sharpness-aware minimization (SAM) technique, utilizing a two-batch strategy to provide a more accurate perturbation estimate and an efficient gradient descent, thereby improving parameter robustness against noise. Notably, DPAdapter can act as a plug-and-play component and be combined with existing DPML algorithms to further improve their performance. Our experiments show that DPAdapter vastly enhances state-of-the-art DPML algorithms, increasing average accuracy from 72.92\% to 77.09\% with a privacy budget of $\epsilon=4$.
Authors: Sotirios Panagiotis Chytas, Vishnu Suresh Lokhande, Peiran Li, Vikas Singh
Abstract: Small sample sizes are common in many disciplines, which necessitates pooling roughly similar datasets across multiple institutions to study weak but relevant associations between images and disease outcomes. Such data often manifest shift/imbalance in covariates (i.e., secondary non-imaging data). Controlling for such nuisance variables is common within standard statistical analysis, but the ideas do not directly apply to overparameterized models. Consequently, recent work has shown how strategies from invariant representation learning provides a meaningful starting point, but the current repertoire of methods is limited to accounting for shifts/imbalances in just a couple of covariates at a time. In this paper, we show how viewing this problem from the perspective of Category theory provides a simple and effective solution that completely avoids elaborate multi-stage training pipelines that would otherwise be needed. We show the effectiveness of this approach via extensive experiments on real datasets. Further, we discuss how this style of formulation offers a unified perspective on at least 5+ distinct problem settings, from self-supervised learning to matching problems in 3D reconstruction.
Authors: Haoyu Chen, Wenbo Li, Jinjin Gu, Jingjing Ren, Haoze Sun, Xueyi Zou, Zhensong Zhang, Youliang Yan, Lei Zhu
Abstract: For image super-resolution (SR), bridging the gap between the performance on synthetic datasets and real-world degradation scenarios remains a challenge. This work introduces a novel "Low-Res Leads the Way" (LWay) training framework, merging Supervised Pre-training with Self-supervised Learning to enhance the adaptability of SR models to real-world images. Our approach utilizes a low-resolution (LR) reconstruction network to extract degradation embeddings from LR images, merging them with super-resolved outputs for LR reconstruction. Leveraging unseen LR images for self-supervised learning guides the model to adapt its modeling space to the target domain, facilitating fine-tuning of SR models without requiring paired high-resolution (HR) images. The integration of Discrete Wavelet Transform (DWT) further refines the focus on high-frequency details. Extensive evaluations show that our method significantly improves the generalization and detail restoration capabilities of SR models on unseen real-world datasets, outperforming existing methods. Our training regime is universally compatible, requiring no network architecture modifications, making it a practical solution for real-world SR applications.
Authors: Xizhi Wang, Yaxiong Wang, Mengjian Li
Abstract: This paper presents a Geometric-Photometric Joint Alignment(GPJA) method, for accurately aligning human expressions by combining geometry and photometric information. Common practices for registering human heads typically involve aligning landmarks with facial template meshes using geometry processing approaches, but often overlook photometric consistency. GPJA overcomes this limitation by leveraging differentiable rendering to align vertices with target expressions, achieving joint alignment in geometry and photometric appearances automatically, without the need for semantic annotation or aligned meshes for training. It features a holistic rendering alignment strategy and a multiscale regularized optimization for robust and fast convergence. The method utilizes derivatives at vertex positions for supervision and employs a gradient-based algorithm which guarantees smoothness and avoids topological defects during the geometry evolution. Experimental results demonstrate faithful alignment under various expressions, surpassing the conventional ICP-based methods and the state-of-the-art deep learning based method. In practical, our method enhances the efficiency of obtaining topology-consistent face models from multi-view stereo facial scanning.
Authors: Ohn Kim, Junwon Seo, Seongyong Ahn, Chong Hui Kim
Abstract: Autonomous off-road navigation requires an accurate semantic understanding of the environment, often converted into a bird's-eye view (BEV) representation for various downstream tasks. While learning-based methods have shown success in generating local semantic terrain maps directly from sensor data, their efficacy in off-road environments is hindered by challenges in accurately representing uncertain terrain features. This paper presents a learning-based fusion method for generating dense terrain classification maps in BEV. By performing LiDAR-image fusion at multiple scales, our approach enhances the accuracy of semantic maps generated from an RGB image and a single-sweep LiDAR scan. Utilizing uncertainty-aware pseudo-labels further enhances the network's ability to learn reliably in off-road environments without requiring precise 3D annotations. By conducting thorough experiments using off-road driving datasets, we demonstrate that our method can improve accuracy in off-road terrains, validating its efficacy in facilitating reliable and safe autonomous navigation in challenging off-road settings.
Authors: HeeSun Bae, Seungjae Shin, Byeonghu Na, Il-Chul Moon
Abstract: For learning with noisy labels, the transition matrix, which explicitly models the relation between noisy label distribution and clean label distribution, has been utilized to achieve the statistical consistency of either the classifier or the risk. Previous researches have focused more on how to estimate this transition matrix well, rather than how to utilize it. We propose good utilization of the transition matrix is crucial and suggest a new utilization method based on resampling, coined RENT. Specifically, we first demonstrate current utilizations can have potential limitations for implementation. As an extension to Reweighting, we suggest the Dirichlet distribution-based per-sample Weight Sampling (DWS) framework, and compare reweighting and resampling under DWS framework. With the analyses from DWS, we propose RENT, a REsampling method with Noise Transition matrix. Empirically, RENT consistently outperforms existing transition matrix utilization methods, which includes reweighting, on various benchmark datasets. Our code is available at \url{https://github.com/BaeHeeSun/RENT}.
Authors: Jiwen Zhang, Jihao Wu, Yihua Teng, Minghui Liao, Nuo Xu, Xiao Xiao, Zhongyu Wei, Duyu Tang
Abstract: Large language model (LLM) leads to a surge of autonomous GUI agents for smartphone, which completes a task triggered by natural language through predicting a sequence of actions of API. Even though the task highly relies on past actions and visual observations, existing studies typical consider little semantic information carried out by intermediate screenshots and screen operations. To address this, this work presents Chain-of-Action-Thought (dubbed CoAT), which takes the description of the previous actions, the current screen, and more importantly the action thinking of what actions should be performed and the outcomes led by the chosen action. We demonstrate that, in a zero-shot setting upon an off-the-shell LLM, CoAT significantly improves the goal progress compared to standard context modeling. To further facilitate the research in this line, we construct a benchmark Android-In-The-Zoo (AitZ), which contains 18,643 screen-action pairs together with chain-of-action-thought annotations. Experiments show that fine-tuning a 200M model on our AitZ dataset achieves on par performance with CogAgent-Chat-18B.
Authors: Mark Karlov, Ali Abedi, Shehroz S. Khan
Abstract: Exercise-based rehabilitation programs have proven to be effective in enhancing the quality of life and reducing mortality and rehospitalization rates. AI-driven virtual rehabilitation, which allows patients to independently complete exercises at home, utilizes AI algorithms to analyze exercise data, providing feedback to patients and updating clinicians on their progress. These programs commonly prescribe a variety of exercise types, leading to a distinct challenge in rehabilitation exercise assessment datasets: while abundant in overall training samples, these datasets often have a limited number of samples for each individual exercise type. This disparity hampers the ability of existing approaches to train generalizable models with such a small sample size per exercise. Addressing this issue, our paper introduces a novel supervised contrastive learning framework with hard and soft negative samples that effectively utilizes the entire dataset to train a single model applicable to all exercise types. This model, with a Spatial-Temporal Graph Convolutional Network (ST-GCN) architecture, demonstrated enhanced generalizability across exercises and a decrease in overall complexity. Through extensive experiments on three publicly available rehabilitation exercise assessment datasets, the University of Idaho-Physical Rehabilitation Movement Data (UI-PRMD), IntelliRehabDS (IRDS), and KInematic assessment of MOvement and clinical scores for remote monitoring of physical REhabilitation (KIMORE), our method has shown to surpass existing methods, setting a new benchmark in rehabilitation exercise assessment accuracy.
Authors: Philipp Hess, Michael Aich, Baoxiang Pan, Niklas Boers
Abstract: Accurate and high-resolution Earth system model (ESM) simulations are essential to assess the ecological and socio-economic impacts of anthropogenic climate change, but are computationally too expensive. Recent machine learning approaches have shown promising results in downscaling ESM simulations, outperforming state-of-the-art statistical approaches. However, existing methods require computationally costly retraining for each ESM and extrapolate poorly to climates unseen during training. We address these shortcomings by learning a consistency model (CM) that efficiently and accurately downscales arbitrary ESM simulations without retraining in a zero-shot manner. Our foundation model approach yields probabilistic downscaled fields at resolution only limited by the observational reference data. We show that the CM outperforms state-of-the-art diffusion models at a fraction of computational cost while maintaining high controllability on the downscaling task. Further, our method generalizes to climate states unseen during training without explicitly formulated physical constraints.
Authors: Kinga Skorupska, Rafa{\l} Stryjek, Izabela Wierzbowska, Piotr Bebas, Maciej Grzeszczuk, Piotr Gago, Jaros{\l}aw Kowalski, Maciej Krzywicki, Jagoda Lazarek, Wies{\l}aw Kope\'c
Abstract: Technology is increasingly used in Nature Reserves and National Parks around the world to support conservation efforts. Endangered species, such as the Eurasian Lynx (Lynx lynx), are monitored by a network of automatic photo traps. Yet, this method produces vast amounts of data, which needs to be prepared, analyzed and interpreted. Therefore, researchers working in this area increasingly need support to process this incoming information. One opportunity is to seek support from volunteer Citizen Scientists who can help label the data, however, it is challenging to retain their interest. Another way is to automate the process with image recognition using convolutional neural networks. During the panel, we will discuss considerations related to nature research and conservation as well as opportunities for the use of Citizen Science and Machine Learning to expedite the process of data preparation, labelling and analysis.
Authors: Ben Pinhasov, Raz Lapid, Rony Ohayon, Moshe Sipper, Yehudit Aperstein
Abstract: We introduce a novel methodology for identifying adversarial attacks on deepfake detectors using eXplainable Artificial Intelligence (XAI). In an era characterized by digital advancement, deepfakes have emerged as a potent tool, creating a demand for efficient detection systems. However, these systems are frequently targeted by adversarial attacks that inhibit their performance. We address this gap, developing a defensible deepfake detector by leveraging the power of XAI. The proposed methodology uses XAI to generate interpretability maps for a given method, providing explicit visualizations of decision-making factors within the AI models. We subsequently employ a pretrained feature extractor that processes both the input image and its corresponding XAI image. The feature embeddings extracted from this process are then used for training a simple yet effective classifier. Our approach contributes not only to the detection of deepfakes but also enhances the understanding of possible adversarial attacks, pinpointing potential vulnerabilities. Furthermore, this approach does not change the performance of the deepfake detector. The paper demonstrates promising results suggesting a potential pathway for future deepfake detection mechanisms. We believe this study will serve as a valuable contribution to the community, sparking much-needed discourse on safeguarding deepfake detectors.
Authors: Haneol Kang, Dong-Wan Choi
Abstract: The stability-plasticity dilemma is a major challenge in continual learning, as it involves balancing the conflicting objectives of maintaining performance on previous tasks while learning new tasks. In this paper, we propose the recall-oriented continual learning framework to address this challenge. Inspired by the human brain's ability to separate the mechanisms responsible for stability and plasticity, our framework consists of a two-level architecture where an inference network effectively acquires new knowledge and a generative network recalls past knowledge when necessary. In particular, to maximize the stability of past knowledge, we investigate the complexity of knowledge depending on different representations, and thereby introducing generative adversarial meta-model (GAMM) that incrementally learns task-specific parameters instead of input data samples of the task. Through our experiments, we show that our framework not only effectively learns new knowledge without any disruption but also achieves high stability of previous knowledge in both task-aware and task-agnostic learning scenarios. Our code is available at: https://github.com/bigdata-inha/recall-oriented-cl-framework.
URLs: https://github.com/bigdata-inha/recall-oriented-cl-framework.
Authors: Chenglei Si, Yanzhe Zhang, Zhengyuan Yang, Ruibo Liu, Diyi Yang
Abstract: Generative AI has made rapid advancements in recent years, achieving unprecedented capabilities in multimodal understanding and code generation. This can enable a new paradigm of front-end development, in which multimodal LLMs might directly convert visual designs into code implementations. In this work, we formalize this as a Design2Code task and conduct comprehensive benchmarking. Specifically, we manually curate a benchmark of 484 diverse real-world webpages as test cases and develop a set of automatic evaluation metrics to assess how well current multimodal LLMs can generate the code implementations that directly render into the given reference webpages, given the screenshots as input. We also complement automatic metrics with comprehensive human evaluations. We develop a suite of multimodal prompting methods and show their effectiveness on GPT-4V and Gemini Pro Vision. We further finetune an open-source Design2Code-18B model that successfully matches the performance of Gemini Pro Vision. Both human evaluation and automatic metrics show that GPT-4V performs the best on this task compared to other models. Moreover, annotators think GPT-4V generated webpages can replace the original reference webpages in 49% of cases in terms of visual appearance and content; and perhaps surprisingly, in 64% of cases GPT-4V generated webpages are considered better than the original reference webpages. Our fine-grained break-down metrics indicate that open-source models mostly lag in recalling visual elements from the input webpages and in generating correct layout designs, while aspects like text content and coloring can be drastically improved with proper finetuning.
Authors: Yicheng Deng, Cheng Sun, Yongqi Sun, Jiahui Zhu
Abstract: 3D human pose estimation from a single image is still a challenging problem despite the large amount of work that has been performed in this field. Generally, most methods directly use neural networks and ignore certain constraints (e.g., reprojection constraints, joint angle, and bone length constraints). While a few methods consider these constraints but train the network separately, they cannot effectively solve the depth ambiguity problem. In this paper, we propose a GAN-based model for 3D human pose estimation, in which a reprojection network is employed to learn the mapping of the distribution from 3D poses to 2D poses, and a discriminator is employed for 2D-3D consistency discrimination. We adopt a novel strategy to synchronously train the generator, the reprojection network and the discriminator. Furthermore, inspired by the typical kinematic chain space (KCS) matrix, we introduce a weighted KCS matrix and take it as one of the discriminator's inputs to impose joint angle and bone length constraints. The experimental results on Human3.6M show that our method significantly outperforms state-of-the-art methods in most cases.
Authors: Emilie Chouzenoux, Marie-Caroline Corbineau, Jean-Christophe Pesquet, Gabriele Scrivanti
Abstract: The joint problem of reconstruction / feature extraction is a challenging task in image processing. It consists in performing, in a joint manner, the restoration of an image and the extraction of its features. In this work, we firstly propose a novel nonsmooth and non-convex variational formulation of the problem. For this purpose, we introduce a versatile generalised Gaussian prior whose parameters, including its exponent, are space-variant. Secondly, we design an alternating proximal-based optimisation algorithm that efficiently exploits the structure of the proposed non-convex objective function. We also analyse the convergence of this algorithm. As shown in numerical experiments conducted on joint deblurring/segmentation tasks, the proposed method provides high-quality results.
Authors: Huixian Cheng, XianFeng Han, Hang Jiang, Dehong He, Guoqiang Xiao
Abstract: Fast and efficient semantic segmentation of large-scale LiDAR point clouds is a fundamental problem in autonomous driving. To achieve this goal, the existing point-based methods mainly choose to adopt Random Sampling strategy to process large-scale point clouds. However, our quantative and qualitative studies have found that Random Sampling may be less suitable for the autonomous driving scenario, since the LiDAR points follow an uneven or even long-tailed distribution across the space, which prevents the model from capturing sufficient information from points in different distance ranges and reduces the model's learning capability. To alleviate this problem, we propose a new Polar Cylinder Balanced Random Sampling method that enables the downsampled point clouds to maintain a more balanced distribution and improve the segmentation performance under different spatial distributions. In addition, a sampling consistency loss is introduced to further improve the segmentation performance and reduce the model's variance under different sampling methods. Extensive experiments confirm that our approach produces excellent performance on both SemanticKITTI and SemanticPOSS benchmarks, achieving a 2.8% and 4.0% improvement, respectively. The source code is available at https://github.com/huixiancheng/PCB-RandNet.
Authors: Anni Tang, Tianyu He, Xu Tan, Jun Ling, Li Song
Abstract: Talking face generation aims at generating photo-realistic video portraits of a target person driven by input audio. Due to its nature of one-to-many mapping from the input audio to the output video (e.g., one speech content may have multiple feasible visual appearances), learning a deterministic mapping like previous works brings ambiguity during training, and thus causes inferior visual results. Although this one-to-many mapping could be alleviated in part by a two-stage framework (i.e., an audio-to-expression model followed by a neural-rendering model), it is still insufficient since the prediction is produced without enough information (e.g., emotions, wrinkles, etc.). In this paper, we propose MemFace to complement the missing information with an implicit memory and an explicit memory that follow the sense of the two stages respectively. More specifically, the implicit memory is employed in the audio-to-expression model to capture high-level semantics in the audio-expression shared space, while the explicit memory is employed in the neural-rendering model to help synthesize pixel-level details. Our experimental results show that our proposed MemFace surpasses all the state-of-the-art results across multiple scenarios consistently and significantly.
Authors: Dengyu Wu, Gaojie Jin, Han Yu, Xinping Yi, Xiaowei Huang
Abstract: Spiking neural network (SNN), next generation of artificial neural network (ANN) that more closely mimic natural neural networks offers promising improvements in computational efficiency. However, current SNN training methodologies predominantly employ a fixed timestep approach, overlooking the potential of dynamic inference in SNN. In this paper, we strengthen the marriage between SNN and event-driven processing with a proposal to consider cutoff in SNN, which can terminate SNN anytime during the inference to achieve efficient inference. Two novel optimisation techniques are presented to achieve inference efficient SNN: a Top-K cutoff and a regularisation. The Top-K cutoff technique optimises the inference of SNN, and the regularisation are proposed to affect the training and construct SNN with optimised performance for cutoff. We conduct an extensive set of experiments on multiple benchmark frame-based datsets, such as Cifar10/100, Tiny-ImageNet and event-based datasets, including CIFAR10-DVS, N-Caltech101 and DVS128 Gesture. The experimental results demonstrate the effectiveness of our techniques in both ANN-to-SNN conversion and direct training, affirming their compatibility and potential benefits in enhancing accuracy and reducing inference timestep when integrated with existing methods. Code available: https://github.com/Dengyu-Wu/SNN-Regularisation-Cutoff
URLs: https://github.com/Dengyu-Wu/SNN-Regularisation-Cutoff
Authors: Jin Fang, Dingfu Zhou, Jingjing Zhao, Chenming Wu, Chulin Tang, Cheng-Zhong Xu, Liangjun Zhang
Abstract: Over the past few years, there has been remarkable progress in research on 3D point clouds and their use in autonomous driving scenarios has become widespread. However, deep learning methods heavily rely on annotated data and often face domain generalization issues. Unlike 2D images whose domains usually pertain to the texture information present in them, the features derived from a 3D point cloud are affected by the distribution of the points. The lack of a 3D domain adaptation benchmark leads to the common practice of training a model on one benchmark (e.g. Waymo) and then assessing it on another dataset (e.g. KITTI). This setting results in two distinct domain gaps: scenarios and sensors, making it difficult to analyze and evaluate the method accurately. To tackle this problem, this paper presents LiDAR Dataset with Cross Sensors (LiDAR-CS Dataset), which contains large-scale annotated LiDAR point cloud under six groups of different sensors but with the same corresponding scenarios, captured from hybrid realistic LiDAR simulator. To our knowledge, LiDAR-CS Dataset is the first dataset that addresses the sensor-related gaps in the domain of 3D object detection in real traffic. Furthermore, we evaluate and analyze the performance using various baseline detectors and demonstrated its potential applications. Project page: https://opendriving.github.io/lidar-cs.
Authors: Shoukun Sun, Min Xian, Fei Xu, Luca Capriotti, Tiankai Yao
Abstract: The click-based interactive segmentation aims to extract the object of interest from an image with the guidance of user clicks. Recent work has achieved great overall performance by employing feedback from the output. However, in most state-of-the-art approaches, 1) the inference stage involves inflexible heuristic rules and requires a separate refinement model, and 2) the number of user clicks and model performance cannot be balanced. To address the challenges, we propose a click-based and mask-guided interactive image segmentation framework containing three novel components: Cascade-Forward Refinement (CFR), Iterative Click Loss (ICL), and SUEM image augmentation. The CFR offers a unified inference framework to generate segmentation results in a coarse-to-fine manner. The proposed ICL allows model training to improve segmentation and reduce user interactions simultaneously. The proposed SUEM augmentation is a comprehensive way to create large and diverse training sets for interactive image segmentation. Extensive experiments demonstrate the state-of-the-art performance of the proposed approach on five public datasets. Remarkably, our model reduces by 33.2\%, and 15.5\% the number of clicks required to surpass an IoU of 0.95 in the previous state-of-the-art approach on the Berkeley and DAVIS sets, respectively.
Authors: Hongkai Zheng, Weili Nie, Arash Vahdat, Anima Anandkumar
Abstract: We propose an efficient approach to train large diffusion models with masked transformers. While masked transformers have been extensively explored for representation learning, their application to generative learning is less explored in the vision domain. Our work is the first to exploit masked training to reduce the training cost of diffusion models significantly. Specifically, we randomly mask out a high proportion (e.g., 50%) of patches in diffused input images during training. For masked training, we introduce an asymmetric encoder-decoder architecture consisting of a transformer encoder that operates only on unmasked patches and a lightweight transformer decoder on full patches. To promote a long-range understanding of full patches, we add an auxiliary task of reconstructing masked patches to the denoising score matching objective that learns the score of unmasked patches. Experiments on ImageNet-256x256 and ImageNet-512x512 show that our approach achieves competitive and even better generative performance than the state-of-the-art Diffusion Transformer (DiT) model, using only around 30% of its original training time. Thus, our method shows a promising way of efficiently training large transformer-based diffusion models without sacrificing the generative performance.
Authors: George Martvel, Ilan Shimshoni, Anna Zamansky
Abstract: The field of animal affective computing is rapidly emerging, and analysis of facial expressions is a crucial aspect. One of the most significant challenges that researchers in the field currently face is the scarcity of high-quality, comprehensive datasets that allow the development of models for facial expressions analysis. One of the possible approaches is the utilisation of facial landmarks, which has been shown for humans and animals. In this paper we present a novel dataset of cat facial images annotated with bounding boxes and 48 facial landmarks grounded in cat facial anatomy. We also introduce a landmark detection convolution neural network-based model which uses a magnifying ensembe method. Our model shows excellent performance on cat faces and is generalizable to human facial landmark detection.
Authors: Georg W\"olflein, Dyke Ferber, Asier Rabasco Meneghetti, Omar S. M. El Nahhas, Daniel Truhn, Zunamys I. Carrero, David J. Harrison, Ognjen Arandjelovi\'c, Jakob N. Kather
Abstract: Stain normalisation is thought to be a crucial preprocessing step in computational pathology pipelines. We question this belief in the context of weakly supervised whole slide image classification, motivated by the emergence of powerful feature extractors trained using self-supervised learning on diverse pathology datasets. To this end, we performed the most comprehensive evaluation of publicly available pathology feature extractors to date, involving more than 8,000 training runs across nine tasks, five datasets, three downstream architectures, and various preprocessing setups. Notably, we find that omitting stain normalisation and image augmentations does not compromise downstream slide-level classification performance, while incurring substantial savings in memory and compute. Using a new evaluation metric that facilitates relative downstream performance comparison, we identify the best publicly available extractors, and show that their latent spaces are remarkably robust to variations in stain and augmentations like rotation. Contrary to previous patch-level benchmarking studies, our approach emphasises clinical relevance by focusing on slide-level biomarker prediction tasks in a weakly supervised setting with external validation cohorts. Our findings stand to streamline digital pathology workflows by minimising preprocessing needs and informing the selection of feature extractors. Code and data are available at https://georg.woelflein.eu/good-features.
Authors: Sergio Naval Marimont, Matthew Baugh, Vasilis Siomos, Christos Tzelepis, Bernhard Kainz, Giacomo Tarroni
Abstract: Unsupervised Anomaly Detection (UAD) techniques aim to identify and localize anomalies without relying on annotations, only leveraging a model trained on a dataset known to be free of anomalies. Diffusion models learn to modify inputs $x$ to increase the probability of it belonging to a desired distribution, i.e., they model the score function $\nabla_x \log p(x)$. Such a score function is potentially relevant for UAD, since $\nabla_x \log p(x)$ is itself a pixel-wise anomaly score. However, diffusion models are trained to invert a corruption process based on Gaussian noise and the learned score function is unlikely to generalize to medical anomalies. This work addresses the problem of how to learn a score function relevant for UAD and proposes DISYRE: Diffusion-Inspired SYnthetic REstoration. We retain the diffusion-like pipeline but replace the Gaussian noise corruption with a gradual, synthetic anomaly corruption so the learned score function generalizes to medical, naturally occurring anomalies. We evaluate DISYRE on three common Brain MRI UAD benchmarks and substantially outperform other methods in two out of the three tasks.
Authors: Chi-Hsi Kung, Chieh-Chi Yang, Pang-Yuan Pao, Shu-Wei Lu, Pin-Lun Chen, Hsin-Cheng Lu, Yi-Ting Chen
Abstract: Intelligent driving systems aim to achieve a zero-collision mobility experience, requiring interdisciplinary efforts to enhance safety performance. This work focuses on risk identification, the process of identifying and analyzing risks stemming from dynamic traffic participants and unexpected events. While significant advances have been made in the community, the current evaluation of different risk identification algorithms uses independent datasets, leading to difficulty in direct comparison and hindering collective progress toward safety performance enhancement. To address this limitation, we introduce \textbf{RiskBench}, a large-scale scenario-based benchmark for risk identification. We design a scenario taxonomy and augmentation pipeline to enable a systematic collection of ground truth risks under diverse scenarios. We assess the ability of ten algorithms to (1) detect and locate risks, (2) anticipate risks, and (3) facilitate decision-making. We conduct extensive experiments and summarize future research on risk identification. Our aim is to encourage collaborative endeavors in achieving a society with zero collisions. We have made our dataset and benchmark toolkit publicly on the project page: https://hcis-lab.github.io/RiskBench/
Authors: Xiaohu Lu, Hayder Radha
Abstract: The adversarial robustness of a model is its ability to resist adversarial attacks in the form of small perturbations to input data. Universal adversarial attack methods such as Fast Sign Gradient Method (FSGM) and Projected Gradient Descend (PGD) are popular for LiDAR object detection, but they are often deficient compared to task-specific adversarial attacks. Additionally, these universal methods typically require unrestricted access to the model's information, which is difficult to obtain in real-world applications. To address these limitations, we present a black-box Scaling Adversarial Robustness (ScAR) method for LiDAR object detection. By analyzing the statistical characteristics of 3D object detection datasets such as KITTI, Waymo, and nuScenes, we have found that the model's prediction is sensitive to scaling of 3D instances. We propose three black-box scaling adversarial attack methods based on the available information: model-aware attack, distribution-aware attack, and blind attack. We also introduce a strategy for generating scaling adversarial examples to improve the model's robustness against these three scaling adversarial attacks. Comparison with other methods on public datasets under different 3D object detection architectures demonstrates the effectiveness of our proposed method. Our code is available at https://github.com/xiaohulugo/ScAR-IROS2023.
Authors: Ji Lin, Hongxu Yin, Wei Ping, Yao Lu, Pavlo Molchanov, Andrew Tao, Huizi Mao, Jan Kautz, Mohammad Shoeybi, Song Han
Abstract: Visual language models (VLMs) rapidly progressed with the recent success of large language models. There have been growing efforts on visual instruction tuning to extend the LLM with visual inputs, but lacks an in-depth study of the visual language pre-training process, where the model learns to perform joint modeling on both modalities. In this work, we examine the design options for VLM pre-training by augmenting LLM towards VLM through step-by-step controllable comparisons. We introduce three main findings: (1) freezing LLMs during pre-training can achieve decent zero-shot performance, but lack in-context learning capability, which requires unfreezing the LLM; (2) interleaved pre-training data is beneficial whereas image-text pairs alone are not optimal; (3) re-blending text-only instruction data to image-text data during instruction fine-tuning not only remedies the degradation of text-only tasks, but also boosts VLM task accuracy. With an enhanced pre-training recipe we build VILA, a Visual Language model family that consistently outperforms the state-of-the-art models, e.g., LLaVA-1.5, across main benchmarks without bells and whistles. Multi-modal pre-training also helps unveil appealing properties of VILA, including multi-image reasoning, enhanced in-context learning, and better world knowledge.
Authors: Jinchao Zhu, Yuxuan Wang, Xiaobing Tu, Siyuan Pan, Pengfei Wan, Gao Huang
Abstract: The Stable Diffusion Model (SDM) is a popular and efficient text-to-image (t2i) generation and image-to-image (i2i) generation model. Although there have been some attempts to reduce sampling steps, model distillation, and network quantization, these previous methods generally retain the original network architecture. Billion scale parameters and high computing requirements make the research of model architecture adjustment scarce. In this work, we first explore the computational redundancy part of the network, and then prune the redundancy blocks of the model and maintain the network performance through a progressive incubation strategy. Secondly, in order to maintaining the model performance, we add cross-layer multi-expert conditional convolution (CLME-Condconv) to the block pruning part to inherit the original convolution parameters. Thirdly, we propose a global-regional interactive (GRI) attention to speed up the computationally intensive attention part. Finally, we use semantic-aware supervision (SAS) to align the outputs of the teacher model and student model at the semantic level. Experiments show that this method can effectively train a lightweight model close to the performance of the original SD model, and effectively improve the model speed under limited resources. Experiments show that the proposed method can effectively train a light-weight model close to the performance of the original SD model, and effectively improve the model speed under limited resources. After acceleration, the UNet part of the model is 22% faster and the overall speed is 19% faster.
Authors: Zhaowei Li, Qi Xu, Dong Zhang, Hang Song, Yiqing Cai, Qi Qi, Ran Zhou, Junting Pan, Zefeng Li, Van Tu Vu, Zhida Huang, Tao Wang
Abstract: Multi-modal large language models have demonstrated impressive performance across various tasks in different modalities. However, existing multi-modal models primarily emphasize capturing global information within each modality while neglecting the importance of perceiving local information across modalities. Consequently, these models lack the ability to effectively understand the fine-grained details of input data, limiting their performance in tasks that require a more nuanced understanding. To address this limitation, there is a compelling need to develop models that enable fine-grained understanding across multiple modalities, thereby enhancing their applicability to a wide range of tasks. In this paper, we propose GroundingGPT, a language enhanced multi-modal grounding model. Beyond capturing global information like other multi-modal models, our proposed model excels at tasks demanding a detailed understanding of local information within the input. It demonstrates precise identification and localization of specific regions in images or moments in videos. To achieve this objective, we design a diversified dataset construction pipeline, resulting in a multi-modal, multi-granularity dataset for model training. The code, dataset, and demo of our model can be found at https: //github.com/lzw-lzw/GroundingGPT.
Authors: Y. Liu, L. Lin, K. K. Y. Wong, X. Tang
Abstract: Weakly-supervised segmentation (WSS) has emerged as a solution to mitigate the conflict between annotation cost and model performance by adopting sparse annotation formats (e.g., point, scribble, block, etc.). Typical approaches attempt to exploit anatomy and topology priors to directly expand sparse annotations into pseudo-labels. However, due to a lack of attention to the ambiguous edges in medical images and insufficient exploration of sparse supervision, existing approaches tend to generate erroneous and overconfident pseudo proposals in noisy regions, leading to cumulative model error and performance degradation. In this work, we propose a novel WSS approach, named ProCNS, encompassing two synergistic modules devised with the principles of progressive prototype calibration and noise suppression. Specifically, we design a Prototype-based Regional Spatial Affinity (PRSA) loss to maximize the pair-wise affinities between spatial and semantic elements, providing our model of interest with more reliable guidance. The affinities are derived from the input images and the prototype-refined predictions. Meanwhile, we propose an Adaptive Noise Perception and Masking (ANPM) module to obtain more enriched and representative prototype representations, which adaptively identifies and masks noisy regions within the pseudo proposals, reducing potential erroneous interference during prototype computation. Furthermore, we generate specialized soft pseudo-labels for the noisy regions identified by ANPM, providing supplementary supervision. Extensive experiments on three medical image segmentation tasks involving different modalities demonstrate that the proposed framework significantly outperforms representative state-of-the-art methods
Authors: Bo Yang, Chen Wang, Xiaoshuang Ma, Beiping Song, Zhuang Liu
Abstract: Effectively and efficiently retrieving images from remote sensing databases is a critical challenge in the realm of remote sensing big data. Utilizing hand-drawn sketches as retrieval inputs offers intuitive and user-friendly advantages, yet the potential of multi-level feature integration from sketches remains underexplored, leading to suboptimal retrieval performance. To address this gap, our study introduces a novel zero-shot, sketch-based retrieval method for remote sensing images, leveraging multi-level feature extraction, self-attention-guided tokenization and filtering, and cross-modality attention update. This approach employs only vision information and does not require semantic knowledge concerning the sketch and image. It starts by employing multi-level self-attention guided feature extraction to tokenize the query sketches, as well as self-attention feature extraction to tokenize the candidate images. It then employs cross-attention mechanisms to establish token correspondence between these two modalities, facilitating the computation of sketch-to-image similarity. Our method significantly outperforms existing sketch-based remote sensing image retrieval techniques, as evidenced by tests on multiple datasets. Notably, it also exhibits robust zero-shot learning capabilities and strong generalizability in handling unseen categories and novel remote sensing data. The method's scalability can be further enhanced by the pre-calculation of retrieval tokens for all candidate images in a database. This research underscores the significant potential of multi-level, attention-guided tokenization in cross-modal remote sensing image retrieval. For broader accessibility and research facilitation, we have made the code and dataset used in this study publicly available online. Code and dataset are available at https://github.com/Snowstormfly/Cross-modal-retrieval-MLAGT.
URLs: https://github.com/Snowstormfly/Cross-modal-retrieval-MLAGT.
Authors: Suraj Mishra, Danny Z. Chen
Abstract: Medical image segmentation using deep neural networks has been highly successful. However, the effectiveness of these networks is often limited by inadequate dense prediction and inability to extract robust features. To achieve refined dense prediction, we propose densely decoded networks (ddn), by selectively introducing 'crutch' network connections. Such 'crutch' connections in each upsampling stage of the network decoder (1) enhance target localization by incorporating high resolution features from the encoder, and (2) improve segmentation by facilitating multi-stage contextual information flow. Further, we present a training strategy based on adaptive deep supervision (ads), which exploits and adapts specific attributes of input dataset, for robust feature extraction. In particular, ads strategically locates and deploys auxiliary supervision, by matching the average input object size with the layer-wise effective receptive fields (lerf) of a network, resulting in a class of ddns. Such inclusion of 'companion objective' from a specific hidden layer, helps the model pay close attention to some distinct input-dependent features, which the network might otherwise 'ignore' during training. Our new networks and training strategy are validated on 4 diverse datasets of different modalities, demonstrating their effectiveness.
Authors: Sachin Chhabra, Hemanth Venkateswara, Baoxin Li
Abstract: In the absence of labeled target data, unsupervised domain adaptation approaches seek to align the marginal distributions of the source and target domains in order to train a classifier for the target. Unsupervised domain alignment procedures are category-agnostic and end up misaligning the categories. We address this problem by deploying a pretrained network to determine accurate labels for the target domain using a multi-stage pseudo-label refinement procedure. The filters are based on the confidence, distance (conformity), and consistency of the pseudo labels. Our results on multiple datasets demonstrate the effectiveness of our simple procedure in comparison with complex state-of-the-art techniques.
Authors: Takuya Ikeda, Sergey Zakharov, Tianyi Ko, Muhammad Zubair Irshad, Robert Lee, Katherine Liu, Rares Ambrus, Koichi Nishiwaki
Abstract: This paper addresses the challenging problem of category-level pose estimation. Current state-of-the-art methods for this task face challenges when dealing with symmetric objects and when attempting to generalize to new environments solely through synthetic data training. In this work, we address these challenges by proposing a probabilistic model that relies on diffusion to estimate dense canonical maps crucial for recovering partial object shapes as well as establishing correspondences essential for pose estimation. Furthermore, we introduce critical components to enhance performance by leveraging the strength of the diffusion models with multi-modal input representations. We demonstrate the effectiveness of our method by testing it on a range of real datasets. Despite being trained solely on our generated synthetic data, our approach achieves state-of-the-art performance and unprecedented generalization qualities, outperforming baselines, even those specifically trained on the target domain.
Authors: Jinsung Jeon, Hyundong Jin, Jonghyun Choi, Sanghyun Hong, Dongeun Lee, Kookjin Lee, Noseong Park
Abstract: A standard practice in developing image recognition models is to train a model on a specific image resolution and then deploy it. However, in real-world inference, models often encounter images different from the training sets in resolution and/or subject to natural variations such as weather changes, noise types and compression artifacts. While traditional solutions involve training multiple models for different resolutions or input variations, these methods are computationally expensive and thus do not scale in practice. To this end, we propose a novel neural network model, parallel-structured and all-component Fourier neural operator (PAC-FNO), that addresses the problem. Unlike conventional feed-forward neural networks, PAC-FNO operates in the frequency domain, allowing it to handle images of varying resolutions within a single model. We also propose a two-stage algorithm for training PAC-FNO with a minimal modification to the original, downstream model. Moreover, the proposed PAC-FNO is ready to work with existing image recognition models. Extensively evaluating methods with seven image recognition benchmarks, we show that the proposed PAC-FNO improves the performance of existing baseline models on images with various resolutions by up to 77.1% and various types of natural variations in the images at inference.
Authors: Zhi-Yi Lin, Bofan Lyu, Judith Cueto Fernandez, Eline van der Kruk, Ajay Seth, Xucong Zhang
Abstract: Accurate 3D kinematics estimation of human body is crucial in various applications for human health and mobility, such as rehabilitation, injury prevention, and diagnosis, as it helps to understand the biomechanical loading experienced during movement. Conventional marker-based motion capture is expensive in terms of financial investment, time, and the expertise required. Moreover, due to the scarcity of datasets with accurate annotations, existing markerless motion capture methods suffer from challenges including unreliable 2D keypoint detection, limited anatomic accuracy, and low generalization capability. In this work, we propose a novel biomechanics-aware network that directly outputs 3D kinematics from two input views with consideration of biomechanical prior and spatio-temporal information. To train the model, we create synthetic dataset ODAH with accurate kinematics annotations generated by aligning the body mesh from the SMPL-X model and a full-body OpenSim skeletal model. Our extensive experiments demonstrate that the proposed approach, only trained on synthetic data, outperforms previous state-of-the-art methods when evaluated across multiple datasets, revealing a promising direction for enhancing video-based human motion capture
Authors: Daan Bon, Gautam Pai, Gijs Bellaard, Olga Mula, Remco Duits
Abstract: The roto-translation group SE2 has been of active interest in image analysis due to methods that lift the image data to multi-orientation representations defined on this Lie group. This has led to impactful applications of crossing-preserving flows for image de-noising, geodesic tracking, and roto-translation equivariant deep learning. In this paper, we develop a computational framework for optimal transportation over Lie groups, with a special focus on SE2. We make several theoretical contributions (generalizable to matrix Lie groups) such as the non-optimality of group actions as transport maps, invariance and equivariance of optimal transport, and the quality of the entropic-regularized optimal transport plan using geodesic distance approximations. We develop a Sinkhorn like algorithm that can be efficiently implemented using fast and accurate distance approximations of the Lie group and GPU-friendly group convolutions. We report valuable advancements in the experiments on 1) image barycentric interpolation, 2) interpolation of planar orientation fields, and 3) Wasserstein gradient flows on SE2. We observe that our framework of lifting images to SE2 and optimal transport with left-invariant anisotropic metrics leads to equivariant transport along dominant contours and salient line structures in the image. This yields sharper and more meaningful interpolations compared to their counterparts on R^2
Authors: Xiaopeng Peng, Erin F. Fleet, Abbie T. Watnik, Grover A. Swartzlander
Abstract: Machine vision is susceptible to laser dazzle, where intense laser light can blind and distort its perception of the environment through oversaturation or permanent damage to sensor pixels. Here we employ a wavefront-coded phase mask to diffuse the energy of laser light and introduce a sandwich generative adversarial network (SGAN) to restore images from complex image degradations, such as varying laser-induced image saturation, mask-induced image blurring, unknown lighting conditions, and various noise corruptions. The SGAN architecture combines discriminative and generative methods by wrapping two GANs around a learnable image deconvolution module. In addition, we make use of Fourier feature representations to reduce the spectral bias of neural networks and improve its learning of high-frequency image details. End-to-end training includes the realistic physics-based synthesis of a large set of training data from publicly available images. We trained the SGAN to suppress the peak laser irradiance as high as $10^6$ times the sensor saturation threshold - the point at which camera sensors may experience damage without the mask. The trained model was evaluated on both a synthetic data set and data collected from the laboratory. The proposed image restoration model quantitatively and qualitatively outperforms state-of-the-art methods for a wide range of scene contents, laser powers, incident laser angles, ambient illumination strengths, and noise characteristics.
Authors: Bin Cao, Jianhao Yuan, Yexin Liu, Jian Li, Shuyang Sun, Jing Liu, Bo Zhao
Abstract: In the rapidly evolving area of image synthesis, a serious challenge is the presence of complex artifacts that compromise perceptual realism of synthetic images. To alleviate artifacts and improve quality of synthetic images, we fine-tune Vision-Language Model (VLM) as artifact classifier to automatically identify and classify a wide range of artifacts and provide supervision for further optimizing generative models. Specifically, we develop a comprehensive artifact taxonomy and construct a dataset of synthetic images with artifact annotations for fine-tuning VLM, named SynArtifact-1K. The fine-tuned VLM exhibits superior ability of identifying artifacts and outperforms the baseline by 25.66%. To our knowledge, this is the first time such end-to-end artifact classification task and solution have been proposed. Finally, we leverage the output of VLM as feedback to refine the generative model for alleviating artifacts. Visualization results and user study demonstrate that the quality of images synthesized by the refined diffusion model has been obviously improved.
Authors: Xinyi Fang, Chak Fong Chong, Kei Long Wong, Yapeng Wang, Wei Ke, Tiankui Zhang, Sio-Kei Im
Abstract: Transferring features learned from natural to medical images for classification is common. However, challenges arise due to the scarcity of certain medical image types and the feature disparities between natural and medical images. Two-step transfer learning has been recognized as a promising solution for this issue. However, choosing an appropriate intermediate domain would be critical in further improving the classification performance. In this work, we explore the effectiveness of using color fundus photographs of the diabetic retina dataset as an intermediate domain for two-step heterogeneous learning (THTL) to classify laryngeal vascular images with nine deep-learning models. Experiment results confirm that although the images in both the intermediate and target domains share vascularized characteristics, the accuracy is drastically reduced compared to one-step transfer learning, where only the last layer is fine-tuned (e.g., ResNet18 drops 14.7%, ResNet50 drops 14.8%). By analyzing the Layer Class Activation Maps (LayerCAM), we uncover a novel finding that the prevalent radial vascular pattern in the intermediate domain prevents learning the features of twisted and tangled vessels that distinguish the malignant class in the target domain. To address the performance drop, we propose the Step-Wise Fine-Tuning (SWFT) method on ResNet in the second step of THTL, resulting in substantial accuracy improvements. Compared to THTL's second step, where only the last layer is fine-tuned, accuracy increases by 26.1% for ResNet18 and 20.4% for ResNet50. Additionally, compared to training from scratch, using ImageNet as the source domain could slightly improve classification performance for laryngeal vascular, but the differences are insignificant.
Authors: Jiahao Zhou, Chen Long, Yue Xie, Jialiang Wang, Boheng Li, Haiping Wang, Zhe Chen, Zhen Dong
Abstract: Developing a unified multi-task foundation model has become a critical challenge in computer vision research. In the current field of 3D computer vision, most datasets only focus on single task, which complicates the concurrent training requirements of various downstream tasks. In this paper, we introduce VEnvision3D, a large 3D synthetic perception dataset for multi-task learning, including depth completion, segmentation, upsampling, place recognition, and 3D reconstruction. Since the data for each task is collected in the same environmental domain, sub-tasks are inherently aligned in terms of the utilized data. Therefore, such a unique attribute can assist in exploring the potential for the multi-task model and even the foundation model without separate training methods. Meanwhile, capitalizing on the advantage of virtual environments being freely editable, we implement some novel settings such as simulating temporal changes in the environment and sampling point clouds on model surfaces. These characteristics enable us to present several new benchmarks. We also perform extensive studies on multi-task end-to-end models, revealing new observations, challenges, and opportunities for future research. Our dataset and code will be open-sourced upon acceptance.
Authors: Xun Lin, Shuai Wang, Rizhao Cai, Yizhong Liu, Ying Fu, Zitong Yu, Wenzhong Tang, Alex Kot
Abstract: Face Anti-Spoofing (FAS) is crucial for securing face recognition systems against presentation attacks. With advancements in sensor manufacture and multi-modal learning techniques, many multi-modal FAS approaches have emerged. However, they face challenges in generalizing to unseen attacks and deployment conditions. These challenges arise from (1) modality unreliability, where some modality sensors like depth and infrared undergo significant domain shifts in varying environments, leading to the spread of unreliable information during cross-modal feature fusion, and (2) modality imbalance, where training overly relies on a dominant modality hinders the convergence of others, reducing effectiveness against attack types that are indistinguishable sorely using the dominant modality. To address modality unreliability, we propose the Uncertainty-Guided Cross-Adapter (U-Adapter) to recognize unreliably detected regions within each modality and suppress the impact of unreliable regions on other modalities. For modality imbalance, we propose a Rebalanced Modality Gradient Modulation (ReGrad) strategy to rebalance the convergence speed of all modalities by adaptively adjusting their gradients. Besides, we provide the first large-scale benchmark for evaluating multi-modal FAS performance under domain generalization scenarios. Extensive experiments demonstrate that our method outperforms state-of-the-art methods. Source code and protocols will be released on https://github.com/OMGGGGG/mmdg.
Authors: Jiakai Sun, Han Jiao, Guangyuan Li, Zhanjie Zhang, Lei Zhao, Wei Xing
Abstract: Constructing photo-realistic Free-Viewpoint Videos (FVVs) of dynamic scenes from multi-view videos remains a challenging endeavor. Despite the remarkable advancements achieved by current neural rendering techniques, these methods generally require complete video sequences for offline training and are not capable of real-time rendering. To address these constraints, we introduce 3DGStream, a method designed for efficient FVV streaming of real-world dynamic scenes. Our method achieves fast on-the-fly per-frame reconstruction within 12 seconds and real-time rendering at 200 FPS. Specifically, we utilize 3D Gaussians (3DGs) to represent the scene. Instead of the na\"ive approach of directly optimizing 3DGs per-frame, we employ a compact Neural Transformation Cache (NTC) to model the translations and rotations of 3DGs, markedly reducing the training time and storage required for each FVV frame. Furthermore, we propose an adaptive 3DG addition strategy to handle emerging objects in dynamic scenes. Experiments demonstrate that 3DGStream achieves competitive performance in terms of rendering speed, image quality, training time, and model storage when compared with state-of-the-art methods.
Authors: Zishi Li, Xiaoxuan Ma, Qiuyan Shang, Wentao Zhu, Hai Ci, Yu Qiao, Yizhou Wang
Abstract: Temporal repetition counting aims to quantify the repeated action cycles within a video. The majority of existing methods rely on the similarity correlation matrix to characterize the repetitiveness of actions, but their scalability is hindered due to the quadratic computational complexity. In this work, we introduce a novel approach that employs an action query representation to localize repeated action cycles with linear computational complexity. Based on this representation, we further develop two key components to tackle the essential challenges of temporal repetition counting. Firstly, to facilitate open-set action counting, we propose the dynamic update scheme on action queries. Unlike static action queries, this approach dynamically embeds video features into action queries, offering a more flexible and generalizable representation. Secondly, to distinguish between actions of interest and background noise actions, we incorporate inter-query contrastive learning to regularize the video representations corresponding to different action queries. As a result, our method significantly outperforms previous works, particularly in terms of long video sequences, unseen actions, and actions at various speeds. On the challenging RepCountA benchmark, we outperform the state-of-the-art method TransRAC by 26.5% in OBO accuracy, with a 22.7% mean error decrease and 94.1% computational burden reduction. Code is available at https://github.com/lizishi/DeTRC.
Authors: Yutian Liu, Wenjun Ke, Jianguo Wei
Abstract: Handwritten mathematical expression recognition (HMER) is challenging in image-to-text tasks due to the complex layouts of mathematical expressions and suffers from problems including over-parsing and under-parsing. To solve these, previous HMER methods improve the attention mechanism by utilizing historical alignment information. However, this approach has limitations in addressing under-parsing since it cannot correct the erroneous attention on image areas that should be parsed at subsequent decoding steps. This faulty attention causes the attention module to incorporate future context into the current decoding step, thereby confusing the alignment process. To address this issue, we propose an attention guidance mechanism to explicitly suppress attention weights in irrelevant areas and enhance the appropriate ones, thereby inhibiting access to information outside the intended context. Depending on the type of attention guidance, we devise two complementary approaches to refine attention weights: self-guidance that coordinates attention of multiple heads and neighbor-guidance that integrates attention from adjacent time steps. Experiments show that our method outperforms existing state-of-the-art methods, achieving expression recognition rates of 60.75% / 61.81% / 63.30% on the CROHME 2014/ 2016/ 2019 datasets.
Authors: Litong Gong, Yiran Zhu, Weijie Li, Xiaoyang Kang, Biao Wang, Tiezheng Ge, Bo Zheng
Abstract: Recently, video generation has achieved significant rapid development based on superior text-to-image generation techniques. In this work, we propose a high fidelity framework for image-to-video generation, named AtomoVideo. Based on multi-granularity image injection, we achieve higher fidelity of the generated video to the given image. In addition, thanks to high quality datasets and training strategies, we achieve greater motion intensity while maintaining superior temporal consistency and stability. Our architecture extends flexibly to the video frame prediction task, enabling long sequence prediction through iterative generation. Furthermore, due to the design of adapter training, our approach can be well combined with existing personalized models and controllable modules. By quantitatively and qualitatively evaluation, AtomoVideo achieves superior results compared to popular methods, more examples can be found on our project website: https://atomo-video.github.io/.
Authors: Tong Zheng, Shusaku Sone, Yoshitaka Ushiku, Yuki Oba, Jiaxin Ma
Abstract: This paper presents a Tri-branch Neural Fusion (TNF) approach designed for classifying multimodal medical images and tabular data. It also introduces two solutions to address the challenge of label inconsistency in multimodal classification. Traditional methods in multi-modality medical data classification often rely on single-label approaches, typically merging features from two distinct input modalities. This becomes problematic when features are mutually exclusive or labels differ across modalities, leading to reduced accuracy. To overcome this, our TNF approach implements a tri-branch framework that manages three separate outputs: one for image modality, another for tabular modality, and a third hybrid output that fuses both image and tabular data. The final decision is made through an ensemble method that integrates likelihoods from all three branches. We validate the effectiveness of TNF through extensive experiments, which illustrate its superiority over traditional fusion and ensemble methods in various convolutional neural networks and transformer-based architectures across multiple datasets.
Authors: Lei Li, Tianfang Zhang, Zhongyu Jiang, Cheng-Yen Yang, Jenq-Neng Hwang, Stefan Oehmcke, Dimitri Pierre Johannes Gominski, Fabian Gieseke, Christian Igel
Abstract: Accurate and consistent methods for counting trees based on remote sensing data are needed to support sustainable forest management, assess climate change mitigation strategies, and build trust in tree carbon credits. Two-dimensional remote sensing imagery primarily shows overstory canopy, and it does not facilitate easy differentiation of individual trees in areas with a dense canopy and does not allow for easy separation of trees when the canopy is dense. We leverage the fusion of three-dimensional LiDAR measurements and 2D imagery to facilitate the accurate counting of trees. We compare a deep learning approach to counting trees in forests using 3D airborne LiDAR data and 2D imagery. The approach is compared with state-of-the-art algorithms, like operating on 3D point cloud and 2D imagery. We empirically evaluate the different methods on the NeonTreeCount data set, which we use to define a tree-counting benchmark. The experiments show that FuseCountNet yields more accurate tree counts.
Authors: Puru Vaish, Shunxin Wang, Nicola Strisciuglio
Abstract: Computer vision models normally witness degraded performance when deployed in real-world scenarios, due to unexpected changes in inputs that were not accounted for during training. Data augmentation is commonly used to address this issue, as it aims to increase data variety and reduce the distribution gap between training and test data. However, common visual augmentations might not guarantee extensive robustness of computer vision models. In this paper, we propose Auxiliary Fourier-basis Augmentation (AFA), a complementary technique targeting augmentation in the frequency domain and filling the augmentation gap left by visual augmentations. We demonstrate the utility of augmentation via Fourier-basis additive noise in a straightforward and efficient adversarial setting. Our results show that AFA benefits the robustness of models against common corruptions, OOD generalization, and consistency of performance of models against increasing perturbations, with negligible deficit to the standard performance of models. It can be seamlessly integrated with other augmentation techniques to further boost performance. Code and models can be found at: https://github.com/nis-research/afa-augment
Authors: Xuweiyi Chen, Tian Xia, Sihan Xu
Abstract: Video Diffusion Models have been developed for video generation, usually integrating text and image conditioning to enhance control over the generated content. Despite the progress, ensuring consistency across frames remains a challenge, particularly when using text prompts as control conditions. To address this problem, we introduce UniCtrl, a novel, plug-and-play method that is universally applicable to improve the spatiotemporal consistency and motion diversity of videos generated by text-to-video models without additional training. UniCtrl ensures semantic consistency across different frames through cross-frame self-attention control, and meanwhile, enhances the motion quality and spatiotemporal consistency through motion injection and spatiotemporal synchronization. Our experimental results demonstrate UniCtrl's efficacy in enhancing various text-to-video models, confirming its effectiveness and universality.
Authors: Aleksandr Dekhovich, David M. J. Tax, Marcel H. F. Sluiter, Miguel A. Bessa
Abstract: Current deep neural networks (DNNs) are overparameterized and use most of their neuronal connections during inference for each task. The human brain, however, developed specialized regions for different tasks and performs inference with a small fraction of its neuronal connections. We propose an iterative pruning strategy introducing a simple importance-score metric that deactivates unimportant connections, tackling overparameterization in DNNs and modulating the firing patterns. The aim is to find the smallest number of connections that is still capable of solving a given task with comparable accuracy, i.e. a simpler subnetwork. We achieve comparable performance for LeNet architectures on MNIST, and significantly higher parameter compression than state-of-the-art algorithms for VGG and ResNet architectures on CIFAR-10/100 and Tiny-ImageNet. Our approach also performs well for the two different optimizers considered -- Adam and SGD. The algorithm is not designed to minimize FLOPs when considering current hardware and software implementations, although it performs reasonably when compared to the state of the art.
Authors: Reza Nasirigerdeh, Reihaneh Torkzadehmahani, Daniel Rueckert, Georgios Kaissis
Abstract: Existing convolutional neural network architectures frequently rely upon batch normalization (BatchNorm) to effectively train the model. BatchNorm, however, performs poorly with small batch sizes, and is inapplicable to differential privacy. To address these limitations, we propose the kernel normalization (KernelNorm) and kernel normalized convolutional layers, and incorporate them into kernel normalized convolutional networks (KNConvNets) as the main building blocks. We implement KNConvNets corresponding to the state-of-the-art ResNets while forgoing the BatchNorm layers. Through extensive experiments, we illustrate that KNConvNets achieve higher or competitive performance compared to the BatchNorm counterparts in image classification and semantic segmentation. They also significantly outperform their batch-independent competitors including those based on layer and group normalization in non-private and differentially private training. Given that, KernelNorm combines the batch-independence property of layer and group normalization with the performance advantage of BatchNorm.
Authors: Nuno Fachada, Diogo de Andrade
Abstract: Synthetic data is essential for assessing clustering techniques, complementing and extending real data, and allowing for more complete coverage of a given problem's space. In turn, synthetic data generators have the potential of creating vast amounts of data -- a crucial activity when real-world data is at premium -- while providing a well-understood generation procedure and an interpretable instrument for methodically investigating cluster analysis algorithms. Here, we present Clugen, a modular procedure for synthetic data generation, capable of creating multidimensional clusters supported by line segments using arbitrary distributions. Clugen is open source, comprehensively unit tested and documented, and is available for the Python, R, Julia, and MATLAB/Octave ecosystems. We demonstrate that our proposal can produce rich and varied results in various dimensions, is fit for use in the assessment of clustering algorithms, and has the potential to be a widely used framework in diverse clustering-related research tasks.
Authors: Huasong Zhou, Xiaowei Xu, Xiaodong Wang, Leon Bevan Bullock
Abstract: Backdoor attack has emerged as a novel and concerning threat to AI security. These attacks involve the training of Deep Neural Network (DNN) on datasets that contain hidden trigger patterns. Although the poisoned model behaves normally on benign samples, it exhibits abnormal behavior on samples containing the trigger pattern. However, most existing backdoor attacks suffer from two significant drawbacks: their trigger patterns are visible and easy to detect by backdoor defense or even human inspection, and their injection process results in the loss of natural sample features and trigger patterns, thereby reducing the attack success rate and model accuracy. In this paper, we propose a novel backdoor attack named SATBA that overcomes these limitations using spatial attention and an U-net based model. The attack process begins by using spatial attention to extract meaningful data features and generate trigger patterns associated with clean images. Then, an U-shaped model is used to embed these trigger patterns into the original data without causing noticeable feature loss. We evaluate our attack on three prominent image classification DNN across three standard datasets. The results demonstrate that SATBA achieves high attack success rate while maintaining robustness against backdoor defenses. Furthermore, we conduct extensive image similarity experiments to emphasize the stealthiness of our attack strategy. Overall, SATBA presents a promising approach to backdoor attack, addressing the shortcomings of previous methods and showcasing its effectiveness in evading detection and maintaining high attack success rate.
Authors: Yihong Xu, Lo\"ick Chambon, \'Eloi Zablocki, Micka\"el Chen, Alexandre Alahi, Matthieu Cord, Patrick P\'erez
Abstract: Motion forecasting is crucial in enabling autonomous vehicles to anticipate the future trajectories of surrounding agents. To do so, it requires solving mapping, detection, tracking, and then forecasting problems, in a multi-step pipeline. In this complex system, advances in conventional forecasting methods have been made using curated data, i.e., with the assumption of perfect maps, detection, and tracking. This paradigm, however, ignores any errors from upstream modules. Meanwhile, an emerging end-to-end paradigm, that tightly integrates the perception and forecasting architectures into joint training, promises to solve this issue. However, the evaluation protocols between the two methods were so far incompatible and their comparison was not possible. In fact, conventional forecasting methods are usually not trained nor tested in real-world pipelines (e.g., with upstream detection, tracking, and mapping modules). In this work, we aim to bring forecasting models closer to the real-world deployment. First, we propose a unified evaluation pipeline for forecasting methods with real-world perception inputs, allowing us to compare conventional and end-to-end methods for the first time. Second, our in-depth study uncovers a substantial performance gap when transitioning from curated to perception-based data. In particular, we show that this gap (1) stems not only from differences in precision but also from the nature of imperfect inputs provided by perception modules, and that (2) is not trivially reduced by simply finetuning on perception outputs. Based on extensive experiments, we provide recommendations for critical areas that require improvement and guidance towards more robust motion forecasting in the real world. The evaluation library for benchmarking models under standardized and practical conditions is provided: \url{https://github.com/valeoai/MFEval}.
Authors: Mengjie Shi, Tom Vercauteren, Wenfeng Xia
Abstract: Photoacoustic (PA) image reconstruction involves acoustic inversion that necessitates the specification of the speed of sound (SoS) within the medium of propagation. Due to the lack of information on the spatial distribution of the SoS within heterogeneous soft tissue, a homogeneous SoS distribution (such as 1540 m/s) is typically assumed in PA image reconstruction, similar to that of ultrasound (US) imaging. Failure to compensate the SoS variations leads to aberration artefacts, deteriorating the image quality. Various methods have been proposed to address this issue, but they usually involve complex hardware and/or time-consuming algorithms, hindering clinical translation. In this work, we introduce a deep learning framework for SoS estimation and subsequent aberration correction in a dual-modal PA/US imaging system exploiting a clinical US probe. As the acquired PA and US images were inherently co-registered, the estimated SoS distribution from US channel data using a deep neural network was incorporated for accurate PA image reconstruction. The framework comprised an initial pre-training stage based on digital phantoms, which was further enhanced through transfer learning using physical phantom data and associated SoS maps obtained from measurements. This framework achieved a root mean square error of 10.2 m/s and 15.2 m/s for SoS estimation on digital and physical phantoms, respectively and structural similarity index measures of up to 0.86 for PA reconstructions as compared to the conventional approach of 0.69. A maximum of 1.2 times improvement in signal-to-noise ratio of PA images was further demonstrated with a human volunteer study. Our results show that the proposed framework could be valuable in various clinical and preclinical applications to enhance PA image reconstruction.
Authors: Jasper W. van der Graaf, Miranda L. van Hooff, Constantinus F. M. Buckens, Matthieu Rutten, Job L. C. van Susante, Robert Jan Kroeze, Marinus de Kleuver, Bram van Ginneken, Nikolas Lessmann
Abstract: This paper presents a large publicly available multi-center lumbar spine magnetic resonance imaging (MRI) dataset with reference segmentations of vertebrae, intervertebral discs (IVDs), and spinal canal. The dataset includes 447 sagittal T1 and T2 MRI series from 218 patients with a history of low back pain and was collected from four different hospitals. An iterative data annotation approach was used by training a segmentation algorithm on a small part of the dataset, enabling semi-automatic segmentation of the remaining images. The algorithm provided an initial segmentation, which was subsequently reviewed, manually corrected, and added to the training data. We provide reference performance values for this baseline algorithm and nnU-Net, which performed comparably. Performance values were computed on a sequestered set of 39 studies with 97 series, which were additionally used to set up a continuous segmentation challenge that allows for a fair comparison of different segmentation algorithms. This study may encourage wider collaboration in the field of spine segmentation and improve the diagnostic value of lumbar spine MRI.
Authors: Junwon Seo, Taekyung Kim, Seongyong Ahn, Kiho Kwak
Abstract: Autonomous navigation in off-road conditions requires an accurate estimation of terrain traversability. However, traversability estimation in unstructured environments is subject to high uncertainty due to the variability of numerous factors that influence vehicle-terrain interaction. Consequently, it is challenging to obtain a generalizable model that can accurately predict traversability in a variety of environments. This paper presents METAVerse, a meta-learning framework for learning a global model that accurately and reliably predicts terrain traversability across diverse environments. We train the traversability prediction network to generate a dense and continuous-valued cost map from a sparse LiDAR point cloud, leveraging vehicle-terrain interaction feedback in a self-supervised manner. Meta-learning is utilized to train a global model with driving data collected from multiple environments, effectively minimizing estimation uncertainty. During deployment, online adaptation is performed to rapidly adapt the network to the local environment by exploiting recent interaction experiences. To conduct a comprehensive evaluation, we collect driving data from various terrains and demonstrate that our method can obtain a global model that minimizes uncertainty. Moreover, by integrating our model with a model predictive controller, we demonstrate that the reduced uncertainty results in safe and stable navigation in unstructured and unknown terrains.
Authors: Yang Miao, Iro Armeni, Marc Pollefeys, Daniel Barath
Abstract: We introduce an online 2D-to-3D semantic instance mapping algorithm aimed at generating comprehensive, accurate, and efficient semantic 3D maps suitable for autonomous agents in unstructured environments. The proposed approach is based on a Voxel-TSDF representation used in recent algorithms. It introduces novel ways of integrating semantic prediction confidence during mapping, producing semantic and instance-consistent 3D regions. Further improvements are achieved by graph optimization-based semantic labeling and instance refinement. The proposed method achieves accuracy superior to the state of the art on public large-scale datasets, improving on a number of widely used metrics. We also highlight a downfall in the evaluation of recent studies: using the ground truth trajectory as input instead of a SLAM-estimated one substantially affects the accuracy, creating a large gap between the reported results and the actual performance on real-world data.
Authors: Christina Kassab, Matias Mattamala, Lintong Zhang, Maurice Fallon
Abstract: Versatile and adaptive semantic understanding would enable autonomous systems to comprehend and interact with their surroundings. Existing fixed-class models limit the adaptability of indoor mobile and assistive autonomous systems. In this work, we introduce LEXIS, a real-time indoor Simultaneous Localization and Mapping (SLAM) system that harnesses the open-vocabulary nature of Large Language Models (LLMs) to create a unified approach to scene understanding and place recognition. The approach first builds a topological SLAM graph of the environment (using visual-inertial odometry) and embeds Contrastive Language-Image Pretraining (CLIP) features in the graph nodes. We use this representation for flexible room classification and segmentation, serving as a basis for room-centric place recognition. This allows loop closure searches to be directed towards semantically relevant places. Our proposed system is evaluated using both public, simulated data and real-world data, covering office and home environments. It successfully categorizes rooms with varying layouts and dimensions and outperforms the state-of-the-art (SOTA). For place recognition and trajectory estimation tasks we achieve equivalent performance to the SOTA, all also utilizing the same pre-trained model. Lastly, we demonstrate the system's potential for planning.
Authors: Luca Scimeca, Alexander Rubinstein, Damien Teney, Seong Joon Oh, Armand Mihai Nicolicioiu, Yoshua Bengio
Abstract: Spurious correlations in the data, where multiple cues are predictive of the target labels, often lead to a phenomenon known as shortcut bias, where a model relies on erroneous, easy-to-learn cues while ignoring reliable ones. In this work, we propose an ensemble diversification framework exploiting Diffusion Probabilistic Models (DPMs) for shortcut bias mitigation. We show that at particular training intervals, DPMs can generate images with novel feature combinations, even when trained on samples displaying correlated input features. We leverage this crucial property to generate synthetic counterfactuals to increase model diversity via ensemble disagreement. We show that DPM-guided diversification is sufficient to remove dependence on primary shortcut cues, without a need for additional supervised signals. We further empirically quantify its efficacy on several diversification objectives, and finally show improved generalization and diversification performance on par with prior work that relies on auxiliary data collection.
Authors: Nikolaos Stathoulopoulos, Mario A. V. Saucedo, Anton Koval, George Nikolakopoulos
Abstract: In the field of resource-constrained robots and the need for effective place recognition in multi-robotic systems, this article introduces RecNet, a novel approach that concurrently addresses both challenges. The core of RecNet's methodology involves a transformative process: it projects 3D point clouds into range images, compresses them using an encoder-decoder framework, and subsequently reconstructs the range image, restoring the original point cloud. Additionally, RecNet utilizes the latent vector extracted from this process for efficient place recognition tasks. This approach not only achieves comparable place recognition results but also maintains a compact representation, suitable for sharing among robots to reconstruct their collective maps. The evaluation of RecNet encompasses an array of metrics, including place recognition performance, the structural similarity of the reconstructed point clouds, and the bandwidth transmission advantages, derived from sharing only the latent vectors. Our proposed approach is assessed using both a publicly available dataset and field experiments$^1$, confirming its efficacy and potential for real-world applications.
Authors: Shih-Yang Liu, Chien-Yi Wang, Hongxu Yin, Pavlo Molchanov, Yu-Chiang Frank Wang, Kwang-Ting Cheng, Min-Hung Chen
Abstract: Among the widely used parameter-efficient finetuning (PEFT) methods, LoRA and its variants have gained considerable popularity because of avoiding additional inference costs. However, there still often exists an accuracy gap between these methods and full fine-tuning (FT). In this work, we first introduce a novel weight decomposition analysis to investigate the inherent differences between FT and LoRA. Aiming to resemble the learning capacity of FT from the findings, we propose Weight-Decomposed LowRank Adaptation (DoRA). DoRA decomposes the pre-trained weight into two components, magnitude and direction, for fine-tuning, specifically employing LoRA for directional updates to efficiently minimize the number of trainable parameters. By employing DoRA, we enhance both the learning capacity and training stability of LoRA while avoiding any additional inference overhead. DoRA consistently outperforms LoRA on fine-tuning LLaMA, LLaVA, and VL-BART on various downstream tasks, such as commonsense reasoning, visual instruction tuning, and image/video-text understanding.
Authors: Muhammad Maaz, Hanoona Rasheed, Abdelrahman Shaker, Salman Khan, Hisham Cholakal, Rao M. Anwer, Tim Baldwin, Michael Felsberg, Fahad S. Khan
Abstract: In pursuit of more inclusive Vision-Language Models (VLMs), this study introduces a Large Multilingual Multimodal Model called PALO. PALO offers visual reasoning capabilities in 10 major languages, including English, Chinese, Hindi, Spanish, French, Arabic, Bengali, Russian, Urdu, and Japanese, that span a total of ~5B people (65% of the world population). Our approach involves a semi-automated translation approach to adapt the multimodal instruction dataset from English to the target languages using a fine-tuned Large Language Model, thereby ensuring high linguistic fidelity while allowing scalability due to minimal manual effort. The incorporation of diverse instruction sets helps us boost overall performance across multiple languages especially those that are underrepresented like Hindi, Arabic, Bengali, and Urdu. The resulting models are trained across three scales (1.7B, 7B and 13B parameters) to show the generalization and scalability where we observe substantial improvements compared to strong baselines. We also propose the first multilingual multimodal benchmark for the forthcoming approaches to evaluate their vision-language reasoning capabilities across languages. Code: https://github.com/mbzuai-oryx/PALO.
Authors: Wen Wen, Mu Li, Yabin Zhang, Yiting Liao, Junlin Li, Li Zhang, Kede Ma
Abstract: Blind video quality assessment (BVQA) plays a pivotal role in evaluating and improving the viewing experience of end-users across a wide range of video-based platforms and services. Contemporary deep learning-based models primarily analyze the video content in its aggressively downsampled format, while being blind to the impact of actual spatial resolution and frame rate on video quality. In this paper, we propose a modular BVQA model, and a method of training it to improve its modularity. Specifically, our model comprises a base quality predictor, a spatial rectifier, and a temporal rectifier, responding to the visual content and distortion, spatial resolution, and frame rate changes on video quality, respectively. During training, spatial and temporal rectifiers are dropped out with some probabilities so as to make the base quality predictor a standalone BVQA model, which should work better with the rectifiers. Extensive experiments on both professionally-generated content and user generated content video databases show that our quality model achieves superior or comparable performance to current methods. Furthermore, the modularity of our model offers a great opportunity to analyze existing video quality databases in terms of their spatial and temporal complexities. Last, our BVQA model is cost-effective to add other quality-relevant video attributes such as dynamic range and color gamut as additional rectifiers.
Authors: Jianan Fan, Dongnan Liu, Hang Chang, Heng Huang, Mei Chen, Weidong Cai
Abstract: Machine learning holds tremendous promise for transforming the fundamental practice of scientific discovery by virtue of its data-driven nature. With the ever-increasing stream of research data collection, it would be appealing to autonomously explore patterns and insights from observational data for discovering novel classes of phenotypes and concepts. However, in the biomedical domain, there are several challenges inherently presented in the cumulated data which hamper the progress of novel class discovery. The non-i.i.d. data distribution accompanied by the severe imbalance among different groups of classes essentially leads to ambiguous and biased semantic representations. In this work, we present a geometry-constrained probabilistic modeling treatment to resolve the identified issues. First, we propose to parameterize the approximated posterior of instance embedding as a marginal von MisesFisher distribution to account for the interference of distributional latent bias. Then, we incorporate a suite of critical geometric properties to impose proper constraints on the layout of constructed embedding space, which in turn minimizes the uncontrollable risk for unknown class learning and structuring. Furthermore, a spectral graph-theoretic method is devised to estimate the number of potential novel classes. It inherits two intriguing merits compared to existent approaches, namely high computational efficiency and flexibility for taxonomy-adaptive estimation. Extensive experiments across various biomedical scenarios substantiate the effectiveness and general applicability of our method.
Authors: Lizhou Fan, Wenyue Hua, Xiang Li, Kaijie Zhu, Mingyu Jin, Lingyao Li, Haoyang Ling, Jinkui Chi, Jindong Wang, Xin Ma, Yongfeng Zhang
Abstract: Understanding the reasoning capabilities of Multimodal Large Language Models (MLLMs) is an important area of research. In this study, we introduce a dynamic benchmark, NPHardEval4V, aimed at addressing the existing gaps in evaluating the pure reasoning abilities of MLLMs. Our benchmark aims to provide a venue to disentangle the effect of various factors such as image recognition and instruction following, from the overall performance of the models, allowing us to focus solely on evaluating their reasoning abilities. It is built by converting textual description of questions from NPHardEval to image representations. Our findings reveal significant discrepancies in reasoning abilities across different models and highlight the relatively weak performance of MLLMs compared to LLMs in terms of reasoning. We also investigate the impact of different prompting styles, including visual, text, and combined visual and text prompts, on the reasoning abilities of MLLMs, demonstrating the different impacts of multimodal inputs in model performance. Unlike traditional benchmarks, which focus primarily on static evaluations, our benchmark will be updated monthly to prevent overfitting and ensure a more authentic and fine-grained evaluation of the models. We believe that this benchmark can aid in understanding and guide the further development of reasoning abilities in MLLMs. The benchmark dataset and code are available at https://github.com/lizhouf/NPHardEval4V
Authors: Damien Teney, Armand Nicolicioiu, Valentin Hartmann, Ehsan Abbasnejad
Abstract: Our understanding of the generalization capabilities of neural networks (NNs) is still incomplete. Prevailing explanations are based on implicit biases of gradient descent (GD) but they cannot account for the capabilities of models from gradient-free methods nor the simplicity bias recently observed in untrained networks. This paper seeks other sources of generalization in NNs. Findings. To understand the inductive biases provided by architectures independently from GD, we examine untrained, random-weight networks. Even simple MLPs show strong inductive biases: uniform sampling in weight space yields a very biased distribution of functions in terms of complexity. But unlike common wisdom, NNs do not have an inherent "simplicity bias". This property depends on components such as ReLUs, residual connections, and layer normalizations. Alternative architectures can be built with a bias for any level of complexity. Transformers also inherit all these properties from their building blocks. Implications. We provide a fresh explanation for the success of deep learning independent from gradient-based training. It points at promising avenues for controlling the solutions implemented by trained models.