检索结果-内蒙古大学图书馆

IEEE International Conference on Multimedia and Expo (ICME)

作者： Dai, Dawei Liu, Yingge Fu, Shiyu Wang, Guoyin Chongqing Univ Posts & Telecommun Chongqing Key Lab Computat Intelligence Key Lab Big Data Intelligent Comp Minist EducKey Lab Cyberspace Big Data Intellige Chongqing 400065 Peoples R China

ISBN: (纸本)9798350390155;9798350390162

Sketch-less facial image retrieval (SLFIR) framework aims to break the barriers that drawing a high-quality facial sketch requires excellent skills and substantial time, it performs the retrieval using a partial sketch with as few strokes as possible. However, such early-stage sketches often contain only local details, resulting in poor retrieval performance. In this study, we propose learning of the representation by fusing the sketches with prior human semantic knowledge to improve the early retrieval performance. Specifically, (1) based on the LAION-Face dataset, a facial language-image pretraining (FLIP) model is constructed to learn the aligned representations of facial image and text;(2) subsequently, using FLIP as the backbone, multiscale features of sketch and text are extracted and fused to learn the efficient representation for the final retrieval. The proposed method achieves state-of-the-art early retrieval performance on all two public datasets and exhibits a good generalization ability in practical testing.

关键词： SBIR SLFIR Multimodal Learning

来源：评论

学校读者我要写书评

暂无评论

Hierarchical feature selection driven by inter-category relevance and semantic independence

引用

PATTERN REcognitION 2025年 161卷

作者： Lin, Zilong Lin, Yaojin Zhang, Yang Zhao, Hong Minnan Normal Univ Sch Comp Sci Zhangzhou 363000 Peoples R China Minnan Normal Univ Key Lab Data Sci & Intelligence Applicat Zhangzhou 363000 Peoples R China

Hierarchical feature selection is crucial in simplifying hierarchical classification and improving time efficiency. The existing methods select features by distinguishing the weight matrix of the current node from its siblings and approximating its parent weight matrix. However, this approach induces a strongly dependent relationship that overuses the weight matrix. Asa result, the model can lose features independent of that system and characterized by semantic independence. In addition, the relevance among categories is ignored considering that by the overuse of the weight matrix, only the correlation of the features and categories is learned. To effectively address these issues, in this work, a method named "Hierarchical Feature Selection Driven by Inter-category Relevance and Semantic Independence (HFSIS)"was proposed to weaken the strongly dependent relationship. HFSIS weakens the strongly dependent relationship by directly multiplying the Spearman correlation matrix with the weight matrix. Thus, the semantic independence features are preserved and the correlation of the features and categories is obtained. Finally, a rich-label information method was used to learn the relevance among the various categories. To demonstrate the superiority of HFSIS, it was compared with eleven state-of-the-art methods in eight hierarchical datasets. It is worth mentioning that HFSIS surpasses other algorithms when using its default parameter settings.

关键词： Hierarchical feature selection Strongly dependent relationship Inter-category relevance Semantic independence

来源：评论

学校读者我要写书评

暂无评论

MultiSum: A Multi -Facet Approach for Extractive Social Summarization Utilizing Semantic and Sociological Relationships 38

MultiSum: A Multi -Facet Approach for Extractive Social Summ...

引用

38th AAAI Conference on Artificial intelligence (AAAI) / 36th Conference on Innovative Applications of Artificial intelligence / 14th Symposium on Educational Advances in Artificial intelligence

作者： Zhao, Tanglong He, Ruifang Xu, Jing Wang, Bo Tianjin Univ Coll Intelligence & Comp Tianjin Peoples R China Tianjin Key Lab Cognit Comp & Applicat Tianjin Peoples R China

ISBN: (纸本)1577358872

Social summarization aims to provide summaries for a large number of social texts (called posts) about a single topic. To extract a summary, both the representation of post and summary selection method are crucial. Previous methods introduce social relation to enhance post embedding to mitigate the sparse representation due to its brief and informal expression. However, they ignore that there are multiple relations between posts. Besides, existing graph -based centrality calculation approaches tend to select posts from one aspect. This leads to facet bias especially when there are multiple viewpoints. In this paper, we propose a model named MultiSum to improve social summarization. Specifically, 1) We use graph convolutional networks to fuse text content with social and semantic relations to improve post representation;2) The similarity between the summary and all aspects is incorporated into the centrality score during the selection phase, encouraging the model to pay attention to different facets. Experimental results on English and Chinese corpora support the effectiveness of this model. Furthermore, external evaluations by human experts and large language models demonstrate the validity of MultiSum in facet coverage and redundancy reduction.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

SeqBMC: Single-cell data processing using iterative block matrix completion algorithm based on matrix factorisation

引用

IET SYSTEMS BIOLOGY 2025年第1期19卷 e70003页

作者： Gong, Lejun Yu, Like Wei, Xinyi Zhou, Shehai Xu, Shuhua Nanjing Univ Posts & Telecommun Sch Comp Sci Jiangsu Key Lab Big Data Secur & Intelligent Proc Nanjing Peoples R China Wenzhou Univ Technol Sch Data Sci & Artificial Intelligence Wenzhou Peoples R China

With the development of high-throughput sequencing technology, the analysis of single-cell RNA sequencing data has become the focus of current research. Matrix analysis and processing of downstream gene expression after preprocessing is a hot topic for researchers. This paper proposed an iterative block matrix completion algorithm, called SeqBMC, based on matrix factorisation. The algorithm is used to complete the missing value of the gene expression matrix caused by the defect of sequencing technology. The gene frequency of the matrix is used to block the matrix, and then the matrix factorisation algorithm is used to complete the small matrix after the block, and then the biological zeros that may exist in the block matrix are retained. Experimental results show that the matrix completion algorithm can significantly improve the classification performance of the gene expression matrix after completion with 86.81% F1 score, which is conducive to the recognition of cell types in sequencing data. Moreover, this completion method can be completed only by the machine learning method without too much prior knowledge related to biology and has good effects. compared with ALRA, SeqBMC increased 5.47% accuracy and 5.03% F1 score. It indicates that SeqBMC has significant advantages in the matrix completion of single-cell RNA sequencing data.

关键词： big data biocomputers biocomputing biology computing data analysis data mining decision making

来源：评论

学校读者我要写书评

暂无评论

Splitting the backbone: A novel hierarchical method for assessing light field image quality

引用

OPTICS AND LASERS IN ENGINEERING 2024年 178卷

作者： Guo, Wenzhong Wang, Hanling Ke, Xiao Fuzhou Univ Coll Comp & Data Sci Fujian Prov Key Lab Networking Comp & Intelligent Fuzhou 350116 Fujian Peoples R China Fuzhou Univ Engn Res Ctr Big Data Intelligence Minist Educ Fuzhou 350116 Peoples R China

The rising popularity of light field imaging underscores the pivotal role of image quality in user experience. However, evaluating the quality of light field images presents significant challenges owing to their highdimensional nature. Current quality assessment methods for light field images predominantly rely on machine learning or statistical analysis, often overlooking the interdependence among pixels. To overcome this limitation, we propose an innovative approach that employs a universal backbone network and introduces a dual-task framework for feature extraction. Specifically, we integrate a staged "primary-secondary" hierarchical evaluation mode into the universal backbone networks, enabling accurate quality score inference while preserving the intrinsic information of the original image. Our proposed approach reduces inference time by over 75% compared to existing methods, simultaneously achieving state-of-the-art results in terms of evaluation metrics. By harnessing the efficiency of neural networks, our framework offers an effective solution for the quality assessment of light field images, providing superior accuracy and speed compared to current methodologies.

关键词： Light field images Image quality assessment Deep learning Multitasking mode

来源：评论

学校读者我要写书评

暂无评论

RTF: Recursive TransFusion for Multi-Modal Image Synthesis

引用

IEEE TRANSACTIONS ON IMAGE PROCESSING 2025年 34卷 1573-1587页

作者： Cao, Bing Qi, Guoliang Zhao, Jiaming Zhu, Pengfei Hu, Qinghua Gao, Xinbo Tianjin Univ Coll Intelligence & Comp Tianjin 300072 Peoples R China Xidian Univ State Key Lab Integrated Serv Networks Xian 710071 Peoples R China Chongqing Univ Posts & Telecommun Key Lab Big Data Intelligent Comp Chongqing 400065 Peoples R China

Multi-modal image synthesis is crucial for obtaining complete modalities due to the imaging restrictions in reality. Current methods, primarily CNN-based models, find it challenging to extract global representations because of local inductive bias, leading to synthetic structure deformation or color distortion. Despite the significant global representation ability of transformer in capturing long-range dependencies, its huge parameter size requires considerable training data. Multi-modal synthesis solely based on one of the two structures makes it hard to extract comprehensive information from each modality with limited data. To tackle this dilemma, we propose a simple yet effective Recursive TransFusion (RTF) framework for multi-modal image synthesis. Specifically, we develop a TransFusion unit to integrate local knowledge extracted from the individual modality by connecting a CNN-based local representation block (LRB) and a transformer-based global fusion block (GFB) via a feature translating gate (FTG). Considering the numerous parameters introduced by the transformer, we further unfold a TransFusion unit with recursive constraint repeatedly, forming recursive TransFusion (RTF), which progressively extracts multi-modal information at different depths. Our RTF remarkably reduces network parameters while maintaining superior performance. Extensive experiments validate our superiority against the competing methods on multiple benchmarks. The source code will be available at https://***/guoliangq/RTF.

关键词： Multi-modal image synthesis TransFusion unit recursive unfolding recursive unfolding global dependencies global dependencies global dependencies

来源：评论

学校读者我要写书评

暂无评论

Communication-efficient clustered federated learning via model distance

引用

MACHINE LEARNING 2024年第6期113卷 3869-3888页

作者： Zhang, Mao Zhang, Tie Cheng, Yifei Bao, Changcun Cao, Haoyu Jiang, Deqiang Xu, Linli Univ Sci & Technol China Sch Comp Sci & Technol Hefei Peoples R China Univ Sci & Technol China Sch Data Sci Hefei Peoples R China State Key Lab Cognit Intelligence Hefei Peoples R China Tencent YouTu Lab Hefei Peoples R China

Clustered Federated Learning (CFL) leverages the differences among data distributions on clients to partition all clients into several clusters for personalized federated training. compared with the conventional federated algorithms such as FedAvg, existing methods for CFL require either more communication costs or multi-stage computation overheads. In this paper, we propose an iterative CFL framework with almost the same communication cost as FedAvg in each round based on a novel model distance. Specifically, the model distance measures the discrepancy between the client model and the cluster model so that we can estimate the cluster identities for clients on the server side. The proposed model distance considers class-wise model dissimilarity, which enables us to apply it to multi-class classification even when the labels are non-iid across clients. To calculate the proposed model distance, we introduce two sampling methods which generate samples from feature distributions approximately without accessing the raw dataset. Experimental results show that our method can achieve superior and comparable performance on non-iid and iid data respectively with less communication cost compared with the baselines.

关键词： Federated learning Client clustering data heterogeneity Model distance

来源：评论

学校读者我要写书评

暂无评论

Public-Private Attributes-Based Variational Adversarial Network for Audio-Visual Cross-Modal Matching

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2024年第9期34卷 8698-8709页

作者： Zheng, Aihua Yuan, Fan Zhang, Haichuan Wang, Jiaxiang Tang, Chao Li, Chenglong Anhui Univ Anhui Prov Key Lab Multimodal Cognit Computat & I Sch Artificial Intelligence Hefei 230601 Peoples R China Anhui Univ Sch Comp Sci & Technol Anhui Prov Key Lab Multimodal Cognit Computat Hefei 230601 Peoples R China Hefei Univ Dept Comp Sci & Technol Hefei 230601 Peoples R China

Existing audio-visual cross-modal matching methods focus on mitigating cross-modal heterogeneity but ignore the impact of intra-class discrepancy of the same identity in different scenarios, which might greatly limit the matching performance. To simultaneously handle both problems of intra-class discrepancy and cross-modal heterogeneity, we propose a novel public-private attributes-based variational adversarial network (P(2)VANet), which captures the consistency within and between classes, for audio-visual cross-modal matching. In particular, P(2)VANet first uses a variational auto-encoder, which captures the inherent global information in diverse scenarios from the hidden variable through reconstruction, to reduce the intraclass discrepancy. Then it integrates a public attributes guidance module to capture the consistency of audio and visual by supervision of the common high-level semantic information to mitigate cross-modal heterogeneity. In addition, P(2)VANet designs a private attributes embedding module to enhance the discriminative features inherent in each class to decrease interclass similarity. Extensive experiments on audio-visual cross-modal matching demonstrate the effectiveness of the proposed approach compared with the state-of-the-art methods.

关键词： Audio-visual cross-modal matching variational adversarial learning public-private attributes metric learning

来源：评论

学校读者我要写书评

暂无评论

Learning a multi-cluster memory prototype for unsupervised video anomaly detection

引用

INFORMATION SCIENCES 2025年 686卷

作者： Wu, Yuntao Zeng, Kun Li, Zuoyong Peng, Zhonghua Chen, Xiaobo Hu, Rong Fujian Univ Technol Sch Comp Sci & Math Fujian Prov Key Lab Big Data Min & Applicat Fuzhou 350118 Peoples R China Minjiang Univ Sch Comp & Big Data Fujian Prov Key Lab Informat Proc & Intelligent Co Fuzhou 350121 Peoples R China Shandong Technol & Business Univ Sch Comp Sci & Technol Yantai 264005 Peoples R China Wuyi Univ Key Lab Cognit Comp & Intelligent Informat Proc Fu Wuyishan 354300 Peoples R China

In recent years, there has been rapid development in video anomaly detection (VAD). The previous methods ignored the differences between normal videos and only emphasized learning the commonalities of normal videos. In order to improve the performance of anomaly detection, we delve into the spatial distribution of normal video features and utilize their differences for clustering, leading to more minor reconstruction errors for normal videos and more significant reconstruction errors for abnormal videos. To achieve this goal, we introduce a Multi-Cluster Memory Prototype framework (MCMP) for VAD, which explores the coarse-grained and finegrained information from video snippet features simultaneously to learn a memory prototype, thereby significantly improving the ability to discriminate abnormal events in complex scenes. First, a video feature clustering method that employs contrastive learning is introduced to group samples sharing similar fine-grained features. Second, the memory mechanism is used to capture the feature distribution of normal samples. Lastly, the Gaussian filter feature transformation method is introduced to make normal and abnormal features more distinguishable. The frame level AUC of MCMP on ShanghaiTech and UCF-Crime benchmark datasets has increased by 1.26% and 0.45% compared to state-of-the-art methods. Our code is publicly available at https:// github .com /WuIkun5658 /MCMP.

关键词： Video anomaly detection Contrastive learning Memory mechanism Clustering Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

A Bi-hemisphere Capsule Network Model for Cross-Subject EEG Emotion Recognition 29th

A Bi-hemisphere Capsule Network Model for Cross-Subject EEG ...

引用

29th International Conference on Neural Information Processing

作者： Luan, Xueying Zhang, Gaoyan Yang, Kai Tianjin Univ Coll Intelligence & Comp Tianjin Key Lab Cognit Comp & Applicat Tianjin Peoples R China

ISBN: (纸本)9789819916443;9789819916450

cognitive neuroscience research has revealed that electroencephalography (EEG) has a strong correlation with human emotions. However, due to the individual differences in EEG signals, the traditional models have the shortcoming of poor generalization ability. Based on the discovery that responses to emotional stimuli in the cerebral hemispheres is asymmetric, in this paper, we propose a Bi-hemispheric Capsule Network (Bi-CapsNet) Model for cross-subject EEG emotion recognition. Specifically, we firstly use a long short term memory (LSTM) layer to learn the asymmetry of emotion expression between the left and right hemispheres of the human brain and the deep representations of all the EEG electrodes'signals in different frequency bands. In order to capture the relationship between EEG channels more detailedly, a special mechanism called routing-by-agreement mechanism has been implemented between LSTM and EmotionCaps. We also use a domain discriminator working corporately with the EmotionCaps to reduce the domain shift between the source domain and the target domain. In addition, we propose a method to reduce the uncertainty of predictions on the target domain data by minimizing the entropy of the prediction posterior. Finally, the cross-subject EEG emotion recognition experiments conducted on two public datasets, SEED and SEED-IV, were that when the length of the EEG data samples is 1 s, the proposed model can obtain better results than most methods on the SEED-IV dataset and also achieves state-of-the-art performance on the SEED dataset.

关键词： EEG emotion recognition Cerebral hemisphere asymmetry Long short term memory (LSTM) Routing-by-agreement mechanism Uncertainty

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：