检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Li, Feiran Xu, Qianqian Bao, Shilong Yang, Zhiyong Cong, Runmin Cao, Xiaochun Huang, Qingming Institute of Information Engineering Chinese Academy of Sciences Beijing China School of Cyber Security University of Chinese Academy of Sciences Beijing China Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China Institute of Information Science Beijing Jiaotong University Beijing China School of Control Science and Engineering Shandong University Jinan China Key Laboratory of Machine Intelligence and System Control Ministry of Education Jinan China School of Cyber Science and Tech. Sun Yat-Sen University Shenzhen Campus China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing China

This paper explores the size-invariance of evaluation metrics in Salient Object Detection (SOD), especially when multiple targets of diverse sizes co-exist in the same image. We observe that current metrics are size-sensitive, where larger objects are focused, and smaller ones tend to be ignored. We argue that the evaluation should be size-invariant because bias based on size is unjustified without additional semantic information. In pursuit of this, we propose a generic approach that evaluates each salient object separately and then combines the results, effectively alleviating the imbalance. We further develop an optimization framework tailored to this goal, achieving considerable improvements in detecting objects of different sizes. Theoretically, we provide evidence supporting the validity of our new metrics and present the generalization analysis of SOD. Extensive experiments demonstrate the effectiveness of our method. The code is available at https://***/Ferry-Li/SI-SOD. Copyright © 2024, The Authors. All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Experimental Study on the Zebra Crossing Behavior of Mixed Bicycles and Pedestrians

SSRN

引用

SSRN 2022年

作者： Ma, Jian Wang, Qiao Chen, Juan Jiang, Rui Song, Weiguo Li, Ruoyu Lian, Liping School of Transportation and Logistics National Engineering Laboratory of Integrated Transportation Big Data Application Technology National United Engineering Laboratory of Integrated and Intelligent Transportation Southwest Jiaotong University Chengdu610031 China Faculty of Geosciences and Environmental Engineering Southwest Jiaotong University Chengdu610031 China Key Laboratory of Transport Industry of Big Data Application Technologies for Comprehensive Transport Ministry of Transport Beijing Jiaotong University Beijing100044 China State Key Laboratory of Fire Science University of Science and Technology of China Hefei230027 China School of Urban Planning and Design Peking University Shenzhen Graduate School School of Architectural Engineering Shenzhen Polytechnic Shenzhen518055 China

Slow transportation mode, consisting of walking and cycling, plays an important role in urban traffic system. Based on the fact that bicycles and pedestrians often share the common road facility, we conducted bicycle-pedestrian mixed flow experiments to investigate the interaction effect among individuals in a 10 m-long, 3 m-wide passageway, mimicking the zebra crossing area. The goal is to explore the influence of bicycles proportion ( p ) on traffic efficiency and investigate the spatial distribution of agents in the mixed flow. Our analyses showed that the increase of bicycle proportion results in longer passing time and less travel time because of large personal space requirements of cyclists. When the bicycle proportion is small or large, the majority pedestrians (bicycles) were divided into several parts by minor bicycles (pedestrians) along the motion direction. We also recorded the following- and lateral- distance between two nearest individuals to calculate the theoretical formula of the passing time for unidirectional bicycle-pedestrian mixed flow based on the existing formula. Comparing the passing time between numerical calculation and experimental data, the effect coefficients for bicycle and pedestrian were obtained. Focusing on the overtaking bicycles, the through gap, lateral and headway distance that decide the overtaking behavior were analyzed. These findings are the first step toward the development of bicycle-pedestrian mixed flow on zebra crossing, which is contribute to provide practical implications for the planning and design of bicycle and pedestrian shared infrastructures. © 2022, The Authors. All rights reserved.

关键词： Travel time

来源：评论

学校读者我要写书评

暂无评论

An Adaptive Framework of Multimodal Emotion Recognition Based on Collaborative Discriminative Learning

An Adaptive Framework of Multimodal Emotion Recognition Base...

引用

International Workshop on Advanced Computational Intelligence (IWACI)

作者： Yadi Wang Xiaoding Guo Yibo Zhang Yiyuan Ren Wendi Huang Zunyan Liu Yuming Feng Xiangguang Dai Wei Zhang Hangjun Che School of Computer and Information Engineering Henan Key Laboratory of Big Data Analysis and Processing Henan University Kaifeng China School of Computer and Information Engineering Henan University Kaifeng China Key Laboratory of Intelligent Information Processing and Control Chongqing Three Gorges University Chongqing China School of Computer Science and Engineering Chongqing Three Gorges University Chongqing China School of Three Gorges Artificial Intelligence Chongqing Three Gorges University Chongqing China College of Electronic and Information Engineering Southwest University Chongqing China

Multimodal emotion recognition has great application scenarios in the field of human-computer interaction, and the main task lies in how to capture and handle the consistency and complementarity of multimodal signals. However, in numerous scenarios, multimodal fusion is significantly challenged by variances in the strengths and weaknesses of different modality signals in carrying emotional information. This paper aims to learn the essential information of strong modalities and inter-modality correlations thus achieving better emotion recognition. We propose an adaptive multimodal emotion recognition framework based on collaborative discriminative learning (AMCDL), a framework for emotion recognition that engages in collaborative learning and adjusts to the information-processing capabilities of various emotion recognition models. AMCDL calculates the inter-modality correlation using Canonical Correlation Analysis (CCA) and uses it as an influencing factor to boost the multiplicative combination method to create a class loss function. When trained on several modalities, the class loss function can integrate complementary information to enable strong modal expression, enhancing emotion perception. On the benchmark dataset CMU-MOSI, AMCDL significantly improves classification accuracy when compared to the state-of-the-art models. The experiments showed the flexibility and generality of AMCDL by applying several emotion recognition models to it and producing better classification results than the original model.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Experimental Study on the Zebra Crossing Behavior of Mixed Bicycles and Pedestrians

SSRN

引用

SSRN 2022年

关键词： Travel time

来源：评论

学校读者我要写书评

暂无评论

Parallel Branches-based Second-Order Transformer for Robust Group Re-identification with Layout-Guided Occlusion Mitigation

引用

Expert Systems with Applications 2025年

作者： Xu Zhang Zhiguang Wu Qinghua Zhang Zuyu Zhang Key Laboratory of Big Data Intelligent Computing Chongqing University of Posts and Telecommunications Chongqing 400065 China Key Laboratory of Tourism Multisource Data Perception and Decision Ministry of Culture and Tourism China Chongqing 400065 China Dept. of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing Chongqing 400065 China

Group Re-identification (GReID) seeks to accurately associate group images with the same members across different cameras, often used in expert and intelligent surveillance systems. However, existing methods mainly focus on the challenges of layout and membership variations while neglecting the occlusion problem caused by layout variations. To address this issue, we propose a novel Parallel Branches-based Second-Order Transformer (PB-SOT) framework, which consists of a Parallel Learning Branch (PLB) module and a Layout-guided Local Feature Sampling (LLFS) module. Specifically, the PLB module designs two group feature transformers with distinct weights to extract global and local features at the group level, respectively. The LLFS module leverages inter-member 2D layout to generate spatial density relationships and performs targeted sampling on local features with the highest density, mitigating the impact of occlusion with layout variations. Experimental evaluations conducted on three available datasets, including RoadGroup, CSG, and SYSU-Group, demonstrate the effectiveness and superiority of our proposed method. https://***/CQRhinoZ/PBSOT .

关键词：

来源：评论

学校读者我要写书评

暂无评论

Incorporating Entity Type Information into Knowledge Representation Learning

Incorporating Entity Type Information into Knowledge Represe...

引用

IEEE International Conference on data Science in Cyberspace (DSC)

作者： Wenyu Huang Guohua Wang Huakui Zhang Feng Chen School of Software Engineering South China University of Technology Key Laboratory of Big Data and Intelligent Robot (South China University of Technology) Ministry of Education Guangzhou China

ISBN: (数字)9781728195582

ISBN: (纸本)9781728195599

Knowledge Representation Learning (KRL), which is also known as Knowledge Embedding, is a very useful method to represent complex relations in knowledge graphs. The low-dimensional representation learned by KRL models makes a contribution to many tasks like recommender system and question answering. Recently, many KRL models are trained using square loss or cross entropy loss based on Closed World Assumption (CWA). Although CWA is an easy way for training, it violates the link prediction task which exploits KRL. To overcome the drawback, in this paper, we introduce a new method, Type-based Prior Possibility Assumption (TPPA). TPPA calculates type based prior possibilities for missing triplets instead of zeros in the training process of KRL to weaken the bad influence of CWA. We compare TPPA with the baseline method CWA in ConvE and TuckER, two common frameworks for knowledge representation learning. The experiment results on FB15k-237 dataset show that TPPA based training method outperforms CWA based training method in link prediction task.

关键词：

来源：评论

学校读者我要写书评

暂无评论

MFCLIP: Multi-modal Fine-grained CLIP for Generalizable Diffusion Face Forgery Detection

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 5888-5903页

作者： Zhang, Yaning Wang, Tianyi Yu, Zitong Gao, Zan Shen, Linlin Chen, Shengyong Qilu University of Technology Shandong Academy of Sciences Faculty of Computer Science and Technology Jinan 250014 China National University of Singapore School of Computing 21 Lower Kent Ridge Rd 119077 Singapore Great Bay University School of Computing and Information Technology Dongguan 523000 China Shenzhen University National Engineering Laboratory for Big Data System Computing Technology Shenzhen 518060 China Qilu University of Technology (Shandong Academy of Sciences) Shandong Artificial Intelligence Institute Jinan 250014 China Tianjin University of Technology Key Laboratory of Computer Vision and System Ministry of Education Tianjin 300384 China Shenzhen University Computer Vision Institute College of Computer Science and Software Engineering Shenzhen 518060 China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen 518129 China Shenzhen University Guangdong Key Laboratory of Intelligent Information Processing Shenzhen 518060 China

The rapid development of photo-realistic face generation methods has raised significant concerns in society and academia, highlighting the urgent need for robust and generalizable face forgery detection (FFD) techniques. Although existing approaches mainly capture face forgery patterns using image modality, other modalities like fine-grained noises and texts are not fully explored, which limits the generalization capability of the model. In addition, most FFD methods tend to identify facial images generated by GAN, but struggle to detect unseen diffusion-synthesized ones. To address the limitations, we aim to leverage the cutting-edge foundation model, contrastive language-image pre-training (CLIP), to achieve generalizable diffusion face forgery detection (DFFD). In this paper, we propose a novel multi-modal fine-grained CLIP (MFCLIP) model, which mines comprehensive and fine-grained forgery traces across image-noise modalities via language-guided face forgery representation learning, to facilitate the advancement of DFFD. Specifically, we devise a fine-grained language encoder (FLE) that extracts fine global language features from hierarchical text prompts. We design a multi-modal vision encoder (MVE) to capture global image forgery embeddings as well as fine-grained noise forgery patterns extracted from the richest patch, and integrate them to mine general visual forgery traces. Moreover, we build an innovative plug-and-play sample pair attention (SPA) method to emphasize relevant negative pairs and suppress irrelevant ones, allowing cross-modality sample pairs to conduct more flexible alignment. Extensive experiments and visualizations show that our model outperforms the state of the arts on different settings like cross-generator, cross-forgery, and cross-dataset evaluations. © 2005-2012 IEEE.

关键词： CLIP Diffusion face forgery detection Image-noise fusion Sample pair attention Transformer

来源：评论

学校读者我要写书评

暂无评论

Fuzzy Granule Density-Based Outlier Detection with Multi-Scale Granular Balls

arXiv

引用

arXiv 2025年

作者： Gao, Can Tan, Xiaofeng Zhou, Jie Ding, Weiping Pedrycz, Witold The College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China Guangdong Key Laboratory of Intelligent Information Processing Shenzhen518060 China The National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen518060 China The School of Artificial Intelligence and Computer Science Nantong University Nantong226019 China The Faculty of Data Science City University of Macau Macau999078 China The Department of Electrical and Computer Engineering University of Alberta Edmonton Canada Systems Research Institute Polish Academy of Sciences Warsaw Poland The Department of Electrical and Computer Engineering King Abdulaziz University Jeddah Saudi Arabia The Department of Computer Engineering Istinye University Istanbul Turkey

Outlier detection refers to the identification of anomalous samples that deviate significantly from the distribution of normal data and has been extensively studied and used in a variety of practical tasks. However, most unsupervised outlier detection methods are carefully designed to detect specified outliers, while real-world data may be entangled with different types of outliers. In this study, we propose a fuzzy rough sets-based multi-scale outlier detection method to identify various types of outliers. Specifically, a novel fuzzy rough sets-based method that integrates relative fuzzy granule density is first introduced to improve the capability of detecting local outliers. Then, a multi-scale view generation method based on granular-ball computing is proposed to collaboratively identify group outliers at different levels of granularity. Moreover, reliable outliers and inliers determined by the three-way decision are used to train a weighted support vector machine to further improve the performance of outlier detection. The proposed method innovatively transforms unsupervised outlier detection into a semi-supervised classification problem and for the first time explores the fuzzy rough sets-based outlier detection from the perspective of multi-scale granular balls, allowing for high adaptability to different types of outliers. Extensive experiments carried out on both artificial and UCI datasets demonstrate that the proposed outlier detection method significantly outperforms the state-of-the-art methods, improving the results by at least 8.48% in terms of the Area Under the ROC Curve (AUROC) index. The source codes are released at https://***/Xiaofeng-Tan/MGBOD. © 2025, CC BY.

关键词： Granulation

来源：评论

学校读者我要写书评

暂无评论

Learning Task-aware Robust Deep Learning Systems

arXiv

引用

arXiv 2020年

作者： Han, Keji Li, Yun Long, Xianzhong Ge, Yao Nanjing University of Posts and Telecommunications China Jiangsu Key Laboratory of Big Data Security & Intelligent Processing

Many works demonstrate that deep learning system is vulnerable to adversarial attack. A deep learning system consists of two parts: the deep learning task and the deep model. Nowadays, most existing works investigate the impact of the deep model on robustness of deep learning systems, ignoring the impact of the learning task. In this paper, we adopt the binary and interval label encoding strategy to redefine the classification task and design corresponding loss to improve robustness of the deep learning system. Our method can be viewed as improving the robustness of deep learning systems from both the learning task and deep model. Experimental results demonstrate that our learning task-aware method is much more robust than traditional classification while retaining the accuracy. © 2020, CC BY.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

HAM:a deep collaborative ranking method incorporating textual information

引用

Frontiers of Information Technology & Electronic Engineering 2020年第8期21卷 1206-1216页

作者： Cheng-wei WANG Teng-fei ZHOU Chen CHEN Tian-lei HU Gang CHEN The Key Laboratory of Big Data Intelligent Computing of Zhejiang Province Hangzhou 310027China State Key Lab of CAD&CG Zhejiang UniversityHangzhou 310027China College of Computer Science and Technology Zhejiang UniversityHangzhou 310027China

The recommendation task with a textual corpus aims to model customer preferences from both user feedback and item textual *** is highly desirable to explore a very deep neural network to capture the complicated nonlinear ***,training a deeper recommender is not as effortless as simply adding layers.A deeper recommender suffers from the gradient vanishing/exploding issue and cannot be easily trained by gradient-based ***,textual descriptions probably contain noisy word *** extracting feature vectors from them can harm the recommender’s *** overcome these difficulties,we propose a new recommendation method named the HighwAy reco Mmender(HAM).HAM explores a highway mechanism to make gradient-based training methods stable.A multi-head attention mechanism is devised to automatically denoise textual ***,a block coordinate descent method is devised to train a deep neural *** studies show that the proposed method outperforms state-of-the-art methods significantly in terms of accuracy.

关键词： Deep learning Recommendation system Highway network Block coordinate descent

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：