检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Thoker, Fida Mohammad Jiang, Letian Zhao, Chen Bagad, Piyush Doughty, Hazel Ghanem, Bernard Snoek, Cees G.M. CEMSE King Abdullah University of Science and Technology Makkah Saudi Arabia VGG University of Oxford Oxford United Kingdom LIACS Leiden University Leiden Netherlands Video & Image Sense Lab University of Amsterdam Amsterdam Netherlands

Continued advances in self-supervised learning have led to significant progress in video representation learning, offering a scalable alternative to supervised approaches by eliminating the need for manual annotations. Despite strong performance on standard action recognition benchmarks, existing video self-supervised learning methods are predominantly evaluated within narrow protocols—typically pre-training on Kinetics-400 and finetuning on similar datasets—limiting our understanding of their generalization capabilities in real-world settings. In this work, we present a comprehensive evaluation of modern video self-supervised learning models, focusing on generalization across four key downstream factors: domain shift, sample efficiency, action granularity, and task diversity. Building on our prior work analyzing benchmark sensitivity in CNN-based contrastive learning, we extend the study to cover current state-of-the-art transformer-based video-only and video-text representation models. Specifically, we benchmark 12 transformer-based methods (7 video-only, 5 video-text) and compare them against 10 CNN-based methods, resulting in over 1100 experiments across 8 datasets and 7 downstream tasks. Our analysis reveals that, despite architectural advancements, transformer-based models remain sensitive to downstream conditions. No single method generalizes consistently across all factors;for instance, video-only transformers are more robust to domain shift, CNN-based models perform better on tasks requiring fine-grained temporal reasoning, and video-text transformers underperform both in several downstream settings despite large-scale pretraining. We also observe that recent transformer-based approaches do not universally outperform earlier methods. These findings provide a detailed understanding of the capabilities and limitations of current video self-supervised learning approaches and establish an extended benchmark for evaluating generalization in video representation

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Many-to-Many Transfer Learning on Motor imagery BCI

Many-to-Many Transfer Learning on Motor Imagery BCI

引用

Decision Aid sciences and Application (DASA), International Conference on

作者： Fatih Altindiş Bülent Yilmaz Marco Congedo Electrical and Electronic Engineering Abdullah Gül University Kayseri Turkey Electrical Engineering Gulf University for Science and Technology Hawally Kuwait Grenoble-Image-sPeach-Signal-Automatics Lab GIPSA-lab University of Grenoble Alpes CNRS Grenoble France

ISBN: (数字)9798350369106

ISBN: (纸本)9798350369113

This paper presents many-to-many domain adaptation strategy, named group learning, for motor imagery brain-computer interfaces (BCIs). Group learning, grounded in Riemannian geometry, simultaneously aligns multiple domains in a unified model, whereas fast alignment approach integrates new, unseen domains without re-estimating alignment matrices for all domains. Group learning creates a single machine learning model using data from previous subjects and/or sessions. Fast alignment utilizes the already trained model for an unseen domain without requiring any additional classifier training. The tests on five publicly available motor imagery databases demonstrate the robustness of group learning against negative learning. The classification accuracy scores of binary and multiclass databases show comparable, if not superior, performance to conventional subject-wise learning method.

关键词： Geometry Training Learning systems Adaptation models Databases Transfer learning Motors Brain modeling Brain-computer interfaces Robustness

来源：评论

学校读者我要写书评

暂无评论

Tiny noise, big mistakes: adversarial perturbations induce errors in brain–computer interface spellers

引用

National science Review 2021年第4期8卷 78-90页

作者： Xiao Zhang Dongrui Wu Lieyun Ding Hanbin Luo Chin-Teng Lin Tzyy-Ping Jung Ricardo Chavarriaga Ministry of Education Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and AutomationHuazhong University of Science and Technology School of Civil Engineering and Mechanics Huazhong University of Science and Technology Centre of Artificial Intelligence Faculty of Engineering and Information Technology University of Technology Sydney Swartz Center for Computational Neuroscience Institute for Neural ComputationUniversity of California San Diego Center for Advanced Neurological Engineering Institute of Engineering in Medicine University of California San Diego ZHAW Data Lab Zürich University of Applied Sciences

An electroencephalogram(EEG)-based brain–computer interface(BCI) speller allows a user to input text to a computer by thought. It is particularly useful to severely disabled individuals, e.g. amyotrophic lateral sclerosis patients, who have no other effective means of communication with another person or a *** studies so far focused on making EEG-based BCI spellers faster and more reliable; however, few have considered their security. This study, for the first time, shows that P300 and steady-state visual evoked potential BCI spellers are very vulnerable, i.e. they can be severely attacked by adversarial perturbations,which are too tiny to be noticed when added to EEG signals, but can mislead the spellers to spell anything the attacker wants. The consequence could range from merely user frustration to severe misdiagnosis in clinical applications. We hope our research can attract more attention to the security of EEG-based BCI spellers, and more broadly, EEG-based BCIs, which has received little attention before.

关键词： electroencephalogram brain-computer interfaces BCI spellers adversarial examples

来源：评论

学校读者我要写书评

暂无评论

SpliceMix: A Cross-scale and Semantic Blending Augmentation Strategy for Multi-label image Classification

引用

IEEE Transactions on Multimedia 2025年

作者： Wang, Lei Zhan, Yibing Ma, Leilei Tao, Dapeng Ding, Liang Gong, Chen Nanjing University of Science and Technology Key Laboratory of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education Jiangsu Key Laboratory of Image and Video Understanding for Social Security School of Computer Science and Engineering Jiangsu Nanjing210094 China JD Explore Academy Beijing100000 China Anhui University School of Computer Science and Technology Anhui Heifei230601 China Yunnan University FIST LAB School of Information Science and Engineering Yunnan Kunming650091 China

Recently, Mix-style data augmentation methods (e.g., Mixup and CutMix) have shown promising performance in various visual tasks. However, these methods are primarily designed for single-label images, ignoring the considerable discrepancies between single- and multi-label images, i.e., a multi-label image involves multiple co-occurred categories and fickle object scales. On the other hand, previous multi-label image classification (MLIC) methods tend to design elaborate models, bringing expensive computation. In this article, we introduce a simple but effective augmentation strategy for multi-label image classification, namely SpliceMix. The "splice"in our method is two-fold: 1) Each mixed image is a splice of several downsampled images in the form of a grid, where the semantics of images attending to mixing are blended without object deficiencies for alleviating co-occurred bias;2) We splice mixed images and the original mini-batch to form a new SpliceMixed mini-batch, which allows an image with different scales to contribute to training together. Furthermore, such splice in our SpliceMixed mini-batch enables interactions between mixed images and original regular images. We also provide a simple and non-parametric extension based on consistency learning (SpliceMix-CL) to show the potential of extending our SpliceMix. Extensive experiments on various tasks demonstrate that only using SpliceMix with a baseline model (e.g., ResNet) achieves better performance than state-of-the-art methods. Moreover, the generalizability of our SpliceMix is further validated by the improvements in current MLIC methods when married with our SpliceMix. The code is available at https://***/zuiran/SpliceMix. © 1999-2012 IEEE.

关键词： image classification

来源：评论

学校读者我要写书评

暂无评论

ViTEraser: Harnessing the Power of Vision Transformers for Scene Text Removal with SegMIM Pretraining

arXiv

引用

arXiv 2023年

作者： Peng, Dezhi Liu, Chongyu Liu, Yuliang Jin, Lianwen South China University of Technology China SCUT-Zhuhai Institute of Modern Industrial Innovation China INTSIG-SCUT Joint Lab of Document Image Analysis and Recognition China Huazhong University of Science and Technology China

Scene text removal (STR) aims at replacing text strokes in natural scenes with visually coherent backgrounds. Recent STR approaches rely on iterative refinements or explicit text masks, resulting in high complexity and sensitivity to the accuracy of text localization. Moreover, most existing STR methods adopt convolutional architectures while the potential of vision Transformers (ViTs) remains largely unexplored. In this paper, we propose a simple-yet-effective ViT-based text eraser, dubbed ViTEraser. Following a concise encoder-decoder framework, ViTEraser can easily incorporate various ViTs to enhance long-range modeling. Specifically, the encoder hierarchically maps the input image into the hidden space through ViT blocks and patch embedding layers, while the decoder gradually upsamples the hidden features to the text-erased image with ViT blocks and patch splitting layers. As ViTEraser implicitly integrates text localization and inpainting, we propose a novel end-to-end pretraining method, termed SegMIM, which focuses the encoder and decoder on the text box segmentation and masked image modeling tasks, respectively. Experimental results demonstrate that ViTEraser with SegMIM achieves state-of-the-art performance on STR by a substantial margin and exhibits strong generalization ability when extended to other tasks, e.g., tampered scene text detection. Furthermore, we comprehensively explore the architecture, pretraining, and scalability of the ViT-based encoder-decoder for STR, which provides deep insights into the application of ViT to the STR field. Code is available at https://***/shannanyinxiang/ViTEraser. Copyright © 2023, The Authors. All rights reserved.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Adaptive and Background-Aware Match for Class-Agnostic Counting

SSRN

引用

SSRN 2023年

作者： Gong, Shenjian Yang, Jian Zhang, Shanshan PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing210094 China Nankai University China

Class-agnostic counting (CAC) aims to count object instances in an image by simply specifying a few exemplar boxes of interest. The key challenge for CAC is how to tailor a desirable interaction between exemplar and query features. Previous CAC methods implement such interaction by solely leveraging standard global feature convolution. We find this interaction leads to under-match caused by intra-class diversity and over-match on background, which severely degrades counting performance. In this work, we propose a novel feature interaction method called Adaptive and Background-Aware Match (ABM) against high intra-class diversity and noisy background. Concretely, given exemplar and query features, we improve the original high-dimensional coupled space match to adaptive orthogonal subspace match (AOM), avoiding under-match caused by intra-class diversity. Moreover, background-specific match (BSM) employs interaction between learnable background prototype and query feature to provide global background priors, making match be aware of background regions. Additionally, we find the high scale variance among different query images leads to bad counting performance for extremely small scale objects. Object scale unify (OSU) is proposed to take the size of the exemplars as the scale prior and resize query images such that all objects are in a uniform average scale. Extensive experiments on FSC-147 show that our method outperforms state-of-the-art approaches. We also conduct extensive ablation studies to demonstrate the effectiveness of each component of our proposed method. The code is publicly available at https://***/Yankeegsj/ABMNet. © 2023, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

A Multi-input Deep Neural Network Framework for Non-invasive Detection of Anemia using Finger Nail images

A Multi-input Deep Neural Network Framework for Non-invasive...

引用

International Conference on Biosignals, images and Instrumentation (ICBSII)

作者： Krithika S Ajay Kumar Reddy Poreddy Thunakala Balakrishna Priyanka Kokil Department of Computer Science and Engineering Amrita Vishwa Vidyapeetham Chennai India Department of Electronics and Communication Engineering Advanced Signal and Image Processing (ASIP) Lab Indian Institute of Information Technology Design and Manufacturing Kancheepuram Chennai

ISBN: (数字)9798350350951

ISBN: (纸本)9798350350968

Anemia, characterized by a deficiency in red blood corpuscles or hemoglobin, poses a significant global health challenge, particularly affecting vulnerable populations. Traditional diagnostic methods often involve invasive procedures, posing challenges in resource-limited settings. This study aims to explore non-invasive anemia detection using fingernail images and convolutional neural networks (CNNs) as a promising alternative to conventional diagnostic approaches. The study utilizes a dataset of fingernail images collected from hospitals in Ghana, comprising both anemic and non-anemic cases. The dataset undergoes preprocessing, including selective enhancement of red components, conversion to the CIElab color space, and feature extraction. A multi-input Deep Neural Network (DNN) framework employing pre-trained CNNs is proposed for anemia classification. The pre-trained CNN architectures include EfficientNet B1, EfficientNet B4, and MobileNet V3. The framework’s performance was assessed using two methodologies: The first involved random shuffling of the dataset, followed by division into training, testing, and validation sets, with evaluation metrics including Accuracy, Precision, F1 scores, and a Confusion Matrix. The second employed five-fold cross-validation, measured using accuracy. The evaluation of the proposed DNN framework using both of the methodologies indicates that EfficientNet B4 achieved the highest testing accuracy (97.87%), precision (97.88%), recall (97.87%), and F1 score (97.88%) and a cross-validation accuracy of 97.37% for the first and second methodologies respectively making it best fit for the proposed DNN framework. The findings demonstrate that the proposed framework yields promising results, especially under the second approach, and opens avenues for further exploration in transfer learning, fine-tuning of deep neural networks for multi-input feature integration, and cross-validation.

关键词： Training Accuracy Transfer learning Sociology Nails Quality assessment Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Graph-based motion prediction for abnormal action detection 2

Graph-based motion prediction for abnormal action detection

引用

2nd ACM International Conference on Multimedia in Asia, MMAsia 2020

作者： Tang, Yao Zhao, Lin Yao, Zhaoliang Gong, Chen Yang, Jian Pca Lab Key Lab of Intelligent Percept. and Syst. for High-Dimensional Information of Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security Nanjing University of Science and Technology China

ISBN: (纸本)9781450383080

Abnormal action detection is the most noteworthy part of anomaly detection, which tries to identify unusual human behaviors in videos. Previous methods typically utilize future frame prediction to detect frames deviating from the normal scenario. While this strategy enjoys success in the accuracy of anomaly detection, critical information such as the cause and location of the abnormality is unable to be acquired. This paper proposes human motion prediction for abnormal action detection. We employ sequence of human poses to represent human motion, and detect irregular behavior by comparing the predicted pose with the actual pose detected in the frame. Hence the proposed method is able to explain why the action is regarded as irregularity and locate where the anomaly happens. Moreover, pose sequence is robust to noise, complex background and small targets in videos. Since posture information is non-Euclidean data, graph convolutional network is adopted for future pose prediction, which not only leads to greater expressive power but also stronger generalization capability. Experiments are conducted both on the widely used anomaly detection dataset ShanghaiTech and our newly proposed dataset NJUST-Anomaly, which mainly contains irregular behaviors happened in the campus. Our dataset expands the existing datasets by giving more abnormal actions attracting public attention in social security, which happen in more complex scenes and dynamic backgrounds. Experimental results on both datasets demonstrate the superiority of our method over the-state-of-the-art methods. The source code and NJUST-Anomaly dataset will be made public at https://***/datangzhengqing/MP-GCN. © 2021 ACM.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Q-Instruct: Improving Low-Level Visual Abilities for Multi-Modality Foundation Models

Q-Instruct: Improving Low-Level Visual Abilities for Multi-M...

引用

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Haoning Wu Zicheng Zhang Erli Zhang Chaofeng Chen Liang Liao Annan Wang Kaixin Xu Chunyi Li Jingwen Hou Guangtao Zhai Geng Xue Wenxiu Sun Qiong Yan Weisi Lin S-Lab for Advanced Intelligence Institute of Image Communication and Network Engineering Shanghai Jiao Tong University Nanyang Technological University Institute for Infocomm Research (IR) Agency for Science Technology and Research (A*STAR) Singapore Sensetime Research

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Multi-modality large language models (MLLMs), as represented by GPT-4V, have introduced a paradigm shift for visual perception and understanding tasks, that a variety of abilities can be achieved within one foundation model. While current MLLMs demonstrate primary low-level visual abilities from the identification of low-level visual attributes (e.g., clarity, brightness) to the evaluation on image quality, there's still an imperative to further improve the accuracy of MLLMs to substantially alleviate human burdens. To address this, we collect the first dataset consisting of human natural language feedback on low-level vision. Each feedback offers a comprehensive description of an image's low-level visual attributes, culminating in an overall quality assessment. The constructed Q-Pathway dataset includes 58K detailed human feedbacks on 18,973 multi-sourced images with diverse low-level appearance. To ensure MLLMs can adeptly handle diverse queries, we further propose a GPT-participated transformation to convert these feedbacks into a rich set of 200K instruction-response pairs, termed Q-Instruct. Experimental results indicate that the Q-Instruct consistently elevates various low-level visual capabilities across multiple base models. We anticipate that our datasets can pave the way for a future that foundation models can assist humans on low-level visual tasks.

关键词： Visualization Accuracy Large language models Natural languages Solids Quality assessment Pattern recognition

来源：评论

学校读者我要写书评

暂无评论

XNLI: Explaining and Diagnosing NLI-based Visual Data Analysis

arXiv

引用

arXiv 2023年

作者： Feng, Yingchaojie Wang, Xingbo Pan, Bo Wong, Kam Kwai Ren, Yi Liu, Shi Yan, Zihan Ma, Yuxin Qu, Huamin Chen, Wei The State Key Lab of CAD & CG Zhejiang University Zhejiang Hangzhou China The Laboratory of Art and Archaeology Image Zhejiang University Ministry of Education China The Hong Kong University of Science and Technology Hong Kong The MIT Media Lab CambridgeMA United States The Department of Computer Science and Engineering Southern University of Science and Technology Guangdong China

Natural language interfaces (NLIs) enable users to flexibly specify analytical intentions in data visualization. However, diagnosing the visualization results without understanding the underlying generation process is challenging. Our research explores how to provide explanations for NLIs to help users locate the problems and further revise the queries. We present XNLI, an explainable NLI system for visual data analysis. The system introduces a Provenance Generator to reveal the detailed process of visual transformations, a suite of interactive widgets to support error adjustments, and a Hint Generator to provide query revision hints based on the analysis of user queries and interactions. Two usage scenarios of XNLI and a user study verify the effectiveness and usability of the system. Results suggest that XNLI can significantly enhance task accuracy without interrupting the NLI-based analysis process. Copyright © 2023, The Authors. All rights reserved.

关键词： Data visualization

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：