IEEE Transactions on Emerging Topics in Computational Intelligence最新文献_第5页

Unified Twin Support Vector Machines for Large-Scale Pattern Classification 大规模模式分类的统一双支持向量机

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2025-12-25 DOI: 10.1109/TETCI.2025.3641799

Sambhav Jain;Reshma Rastogi

{"title":"Unified Twin Support Vector Machines for Large-Scale Pattern Classification","authors":"Sambhav Jain;Reshma Rastogi","doi":"10.1109/TETCI.2025.3641799","DOIUrl":"https://doi.org/10.1109/TETCI.2025.3641799","url":null,"abstract":"Over the years, researchers have proposed numerous Twin Support Vector Machines (TSVM) variants aimed at addressing diverse challenges. These variants encompass sparse TSVM models, robust TSVM models as well as those designed to exhibit noise resilience through the modification of loss functions, among other enhancements. However, there is a lack of an approach that can address multiple problems comprehensively. This motivates us to introduce a modular TSVM framework capable of addressing many challenges in a unified sense. Through this paper, we propose Unified-TSVM (Uni-TSVM), which enables efficient handling of any convex as well as non-convex, Least Squares type Difference of Convex (LS-DC) loss functions. We expand the list of LS-DC loss functions by adding a noise-resilient loss and introduce a versatile trick capable of bounding any loss function. The unified framework leverages low-rank property to introduce sparsity into Unified-TSVM, thus accelerating computation for large-scale applications. Moreover, by utilising matrix calculations, the Unified-TSVM framework facilitates explicit classification across binary, multiclass, and multilabel domains while also enhancing computational efficiency. Additionally, we introduce a thresholding mechanism to avoid training the second hyperplane, further reducing computational complexity. Experiments on many large binary, multiclass and multilabel datasets, with and without noise, demonstrate the effectiveness of the proposed approach.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"2049-2060"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148510431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

AGCTO: Attributed Graph Clustering With Transitive Order Convolutional Autoencoder 具有传递阶卷积自编码器的属性图聚类

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2025-01-16 DOI: 10.1109/TETCI.2025.3526281

Ying Xie;Jixiang Wang;Weihua Liu;Hao Wang;Yushan Pan;Lijie Wen;Rongbin Xu;Weiping Ding;Yun Yang

{"title":"AGCTO: Attributed Graph Clustering With Transitive Order Convolutional Autoencoder","authors":"Ying Xie;Jixiang Wang;Weihua Liu;Hao Wang;Yushan Pan;Lijie Wen;Rongbin Xu;Weiping Ding;Yun Yang","doi":"10.1109/TETCI.2025.3526281","DOIUrl":"https://doi.org/10.1109/TETCI.2025.3526281","url":null,"abstract":"Attributed graphs, which associate vertices or edges with attributes, are essential in applications like social network analysis, recommendation systems, bioinformatics, and healthcare analysis. However, their high-dimensional and non-Euclidean nature complicates tasks such as vertex classification, community detection, graph visualization, and graph embedding. Existing techniques often inadequately integrate structural and attribute data, and reconstruction-centric models can compromise graph topology. Additionally, many methods require extensive manual hyperparameter tuning. To address these challenges, we introduce <underline>Attributed <underline>Graph <underline>Clustering with <underline>Transitive <underline>Order convolutional autoencoder (AGCTO). AGCTO uses a Graph Transitive Convolutional AutoEncoder (GTCAE) framework that integrates structural and attribute information. A key innovation is the Graph ORder Distance (GORD), capturing topological relationships to enhance clustering performance. AGCTO's unified loss function combines GTCAE loss on the original graph, GTCAE loss on a simplified graph via GORD, and a topology-preserving loss. This ensures effective fusion of attribute and structure data and preserves graph topology. Evaluations on four real-world and two synthetic attributed graphs show AGCTO's superiority in clustering accuracy, normalized mutual information, F1-score, and Q-modularity. AGCTO offers a robust solution for attributed graph clustering, maintaining graph topology and delivering superior performance.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"1111-1122"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148511182","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Cross-Domain Self-Supervised Learning for Histopathological Images With Multi-Granularity Constraints 基于多粒度约束的组织病理学图像的跨域自监督学习

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2025-11-24 DOI: 10.1109/TETCI.2025.3631709

Yating Huang;Dougal Ferguson;Ashwin Sachdeva;Peter Gardner;Hujun Yin

{"title":"Cross-Domain Self-Supervised Learning for Histopathological Images With Multi-Granularity Constraints","authors":"Yating Huang;Dougal Ferguson;Ashwin Sachdeva;Peter Gardner;Hujun Yin","doi":"10.1109/TETCI.2025.3631709","DOIUrl":"https://doi.org/10.1109/TETCI.2025.3631709","url":null,"abstract":"In pathological image analysis, the scarcity of labeled data and the difficulty of producing scores of annotated images to high-level details hinder the development of large, high-quality datasets, which in turn affects the advancement of deep learning-based diagnostic tools. The lack of annotations can be addressed to some extent by self-supervised learning methods, leveraging the production of general representations from unlabeled data. However, in such deployment, self-supervised learning methods and general machine learning models suffer from significant performance degradation when transferring to a dataset from an unseen clinic or device, due to domain shifts in data distribution between different datasets across clinics. Such domain shifts can cause high inter-class similarities and low cross-domain correlations. Previous domain generalization methods are often susceptible to under-generalization and over-generalization, especially in handling histopathological images. In this paper, we propose a novel cross-domain self-supervised learning method based on multi-granularity constraints. It combines pathology-specific imaging mechanisms to bound the generalized features of histopathological images in a stable representation space and to extend the generalization distance as much as possible under an adversarial constraint for preventing over-generalization. The effectiveness of the proposed approach against the state-of-the-art methods has been demonstrated in a variety of analysis tasks on four independent prostate histopathological imagery datasets. The proposed method achieves an improved generalization in fine-tuned classification and segmentation using limited annotated samples.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"1352-1364"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148511355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Study on Privacy Preserving and Accuracy of Negative Survey Under Non-Exclusive Choice Selection 非排他性选择下负面调查的隐私保护与准确性研究

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2025-10-31 DOI: 10.1109/TETCI.2025.3619772

Ran Liu;Chi-Hung Chi;Borui Cai;Guangyan Huang;Yankin Chi;Raymond Wong;Yanfeng Shu;Jing He

{"title":"Study on Privacy Preserving and Accuracy of Negative Survey Under Non-Exclusive Choice Selection","authors":"Ran Liu;Chi-Hung Chi;Borui Cai;Guangyan Huang;Yankin Chi;Raymond Wong;Yanfeng Shu;Jing He","doi":"10.1109/TETCI.2025.3619772","DOIUrl":"https://doi.org/10.1109/TETCI.2025.3619772","url":null,"abstract":"Negative survey aims for a cost-effective privacy preserving mechanism for multiple choice question-answering. In most existing work, one important assumption is the exclusivity of answers. However, there are many situations where all the selectable answers are not exclusive - they just have different degrees of an interviewee’s preference. This results in significant distortion of the accuracy of the reconstructed distribution. In this paper, we propose to extend current negative survey models to address surveys with non-exclusive selectable answers. Based on our new model, we investigate the relationship between the accuracy of the reconstructed distribution and the loss of personal privacy in detail. We show that in many cases, the accuracy of the reconstructed distribution can be improved substantially with acceptable slight loss in personal privacy if the interviewee is willing to indicate his preference degree to the negative answer he chooses. Furthermore, when the exclusivity of the positive answer goes below a certain threshold, it is actually possible for both the accuracy and privacy of the survey to be improved together. It is because the selected negative answer with higher preference degree might be more preferable than the positive answer with lower preference degree in the distribution reconstruction.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"1945-1958"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148511940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An Alliance for Uncertainty Quantification: Type-2 Fuzzy Logic Systems With Conformal Prediction 不确定性量化联盟：具有保形预测的2型模糊逻辑系统

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2026-01-22 DOI: 10.1109/TETCI.2025.3647649

Saleh Msaddi;Tufan Kumbasar

{"title":"An Alliance for Uncertainty Quantification: Type-2 Fuzzy Logic Systems With Conformal Prediction","authors":"Saleh Msaddi;Tufan Kumbasar","doi":"10.1109/TETCI.2025.3647649","DOIUrl":"https://doi.org/10.1109/TETCI.2025.3647649","url":null,"abstract":"Uncertainty Quantification (UQ) plays a crucial role in high-risk applications, where reliable and precise Predictive Intervals (PIs) are essential for informed decision-making. This study introduces an Alliance for Uncertainty Quantification (UQ) that integrates IT2-FLSs with Conformal Prediction (CP) to generate High-Quality (HQ) PIs. The Alliance leverages the inherent ability of IT2-FLSs to model uncertainties through their structure, combined with CP as an additional layer to provide statistical guarantees that the generated PIs achieve marginal coverage for a desired confidence level. We describe three CP methods used in the Alliance to transform the PIs of any IT2-FLS into HQ-PIs, ensuring tightness and compactness of the PIs. We investigate the synergy between the CP methods and the IT2-FLS structure, particularly the Center of Sets Calculation Methods (CSCM), and demonstrate that the choice of both the CP method and CSCM significantly impacts the performance of the Alliance. We validate our proposed framework on publicly available datasets. The comparative results indicate that the Alliance outperforms standalone IT2-FLS and CP models. The results of the paper underscore the effectiveness of the Alliance in accurately quantifying uncertainties and generating HQ-PIs, making it a robust solution for high-risk applications requiring reliable and precise predictions.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"2022-2032"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148511945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Unified Brain Network Representation Learning via Adaptive Multimodal Fusion for Alzheimer’s Disease Analysis 基于自适应多模态融合的统一脑网络表征学习用于阿尔茨海默病分析

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2025-10-28 DOI: 10.1109/TETCI.2025.3619586

Hongjie Jiang;Yongcheng Zong;Guoheng Huang;Sibo Qiao;Shanshan Wang;Peipeng Liang;Michael Kwok-Po Ng;Shuqiang Wang

{"title":"Unified Brain Network Representation Learning via Adaptive Multimodal Fusion for Alzheimer’s Disease Analysis","authors":"Hongjie Jiang;Yongcheng Zong;Guoheng Huang;Sibo Qiao;Shanshan Wang;Peipeng Liang;Michael Kwok-Po Ng;Shuqiang Wang","doi":"10.1109/TETCI.2025.3619586","DOIUrl":"https://doi.org/10.1109/TETCI.2025.3619586","url":null,"abstract":"The fusion of structural and functional brain network analysis has been widely applied in the analysis of various brain diseases, especially in identifying the progression stages of Alzheimer’s disease. However, most existing multimodal brain network analysis methods separately construct functional and structural networks, making it difficult to incorporate complementary information from several modalities. To tackle this difficulty, we propose AMFusion(Adaptive Multimodal Fusion), a unified brain network construction framework that jointly learns from functional and structural images, thereby efficiently resolving connectivity and node feature learning problems. Our approach begins with a designed quantization encoder to extract structural features from DTI, while dynamic functional connectivity (FC) is constructed from fMRI. Subsequently, a multilevel brain network fusion module is employed to learn and integrate brain connections, considering interactions across different temporal and spatial scales. Finally, a classifier guides further optimization of the brain network, facilitating multi-class disease classification. The performance of the proposed AMFusion was validated using the authentic dataset from ADNI, and experimental results demonstrated the effectiveness of AMFusion. By considering both connection patterns and node features, our method overcomes the limitations of existing approaches, providing a more comprehensive and robust framework for analyzing and diagnosing AD.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"2009-2021"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148504255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deep Illumination Invariance Feature Enhancement and Trilateral Feature Aggregation for Stereo Matching Network 基于深度光照不变性特征增强和三边特征聚合的立体匹配网络

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2026-02-18 DOI: 10.1109/TETCI.2026.3663502

Guowei An;Yaonan Wang;Qing Zhu;Kai Zeng;Xiaofang Yuan;Xuebing Liu

{"title":"Deep Illumination Invariance Feature Enhancement and Trilateral Feature Aggregation for Stereo Matching Network","authors":"Guowei An;Yaonan Wang;Qing Zhu;Kai Zeng;Xiaofang Yuan;Xuebing Liu","doi":"10.1109/TETCI.2026.3663502","DOIUrl":"https://doi.org/10.1109/TETCI.2026.3663502","url":null,"abstract":"Stereo matching is an important part of robot vision intelligent environment perception. Currently, reflective regions and large textureless regions, repetitive texture regions, and depth discontinuity regions are still difficult and error-prone for stereo matching tasks. We propose a novel stereo matching network that can effectively solve the above problems. First, to address the difficulty in reflective regions and large textureless regions, we propose the illumination invariance feature enhancing module which can promote the feature extraction network to extract more sufficient features in reflective regions and large textureless regions. The illumination invariance feature enhancing module is a differentiable transformation of the Census-based matching cost computation and can be combined with multiple stereo networks to improve their performance in difficult regions. Then to address the difficulty in repetitive texture regions and depth discontinuity regions, we propose the differentiable trilateral feature aggregation module to make full use of the prior geometric knowledge of color similarity, spatial distance, and edge strength in the scene to enhance the aggregation ability of cost agregation network in depth discontinuity regions and repetitive texture regions. The proposed network is end-to-end and is verified by extensive experiments. The results show that the proposed method can effectively solve the problems of stereo matching in the above difficult regions and has achieved the competitive performance on Scene Flow datasets, KITTI 2012 datasets, and KITTI 2015 datasets.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"1376-1388"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148508986","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Genetic Programming for Malignancy Diagnosis From Breast Cancer Histopathological Images: A Feature Learning Approach 从乳腺癌组织病理图像中诊断恶性肿瘤的遗传规划：一种特征学习方法

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2025-01-16 DOI: 10.1109/TETCI.2024.3523769

Qurrat Ul Ain;Harith Al-Sahaf;Bing Xue;Mengjie Zhang

{"title":"Genetic Programming for Malignancy Diagnosis From Breast Cancer Histopathological Images: A Feature Learning Approach","authors":"Qurrat Ul Ain;Harith Al-Sahaf;Bing Xue;Mengjie Zhang","doi":"10.1109/TETCI.2024.3523769","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3523769","url":null,"abstract":"Identifying breast cancer using histopathological images is crucial for early detection and treatment of breast cancer. Histopathological images suffer from a high inter-class and intra-class variability, making breast cancer identification a challenging task. Integration of well-developed feature descriptors into learning algorithms can enhance the automatic extraction of high-level features from these images. With its flexible representation and global search abilities, Genetic Programming (GP) is a good learning algorithm to potentially accomplish this goal. This paper proposes a new GP-based feature learning method for automatically selecting and combining various image descriptors to detect breast cancer from histopathological images, which is an emerging topic in computational intelligence. In the new approach, various global features can be learned for the task of breast cancer image classification and it is capable of automatically evolving solutions. A significant improvement in the classification performance comes from the new approach compared to the existing methods on real-world histopathological image dataset. Taking a closer look at the evolved solutions helps identify the most effective feature descriptors for breast cancer image classification. Unlike the black-box models evolved in existing methods, the proposed method evolves models/solutions that can assist dermatologists in making diagnoses by identifying breast cancer characteristics captured by the feature descriptors that are automatically selected during the evolutionary process.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"1123-1136"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148510718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An Autonomous Hexapod Robot With CPG NN- Based Terrain Adaptation Gait and Evolutionary Interpretable Fuzzy Wall-Following Control 基于CPG神经网络的地形自适应步态和进化可解释模糊跟随控制的自主六足机器人

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2025-12-29 DOI: 10.1109/TETCI.2025.3645588

Chia-Feng Juang;Bing-Ruei He;Cheng-Chieh Tsai

{"title":"An Autonomous Hexapod Robot With CPG NN- Based Terrain Adaptation Gait and Evolutionary Interpretable Fuzzy Wall-Following Control","authors":"Chia-Feng Juang;Bing-Ruei He;Cheng-Chieh Tsai","doi":"10.1109/TETCI.2025.3645588","DOIUrl":"https://doi.org/10.1109/TETCI.2025.3645588","url":null,"abstract":"This paper proposes an autonomous hexapod robot with the abilities of gait adaptation to uneven terrain and walking along a wall. In lower-level gait adaptation control, the forward-walking gait of the robot is generated using central pattern generator (CPG) neural networks. This paper proposes a memory-based, proportional-derivative (m-PD) controller to control the CPG waves of knee joints to maintain body balance on uneven terrain. In higher-level control, this paper employs an interpretable fuzzy system (IFS) to control the CPG waves of hip joints to execute the wall-following task. The IFS considers not only control performance but also model interpretability. To improve learning performance, this paper proposes a multiobjective evolutionary IFS through rule and fuzzy-set pruning (MEIFS-RFP) to optimize the IFS structure and parameters. Simulations and comparisons with different control and learning methods are performed to show the advantages of the proposed control approach. Experiments on a real hexapod robot show the effectiveness of the control approach in real environments.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"1256-1270"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148511183","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multi-Perceptual Framework for Anomaly Detection 异常检测的多感知框架

IF 6 3区计算机科学

IEEE Transactions on Emerging Topics in Computational Intelligence Pub Date : 2026-04-01 Epub Date: 2026-02-02 DOI: 10.1109/TETCI.2026.3657946

Chaobo Li;Hongjun Li;Guoan Zhang

{"title":"Multi-Perceptual Framework for Anomaly Detection","authors":"Chaobo Li;Hongjun Li;Guoan Zhang","doi":"10.1109/TETCI.2026.3657946","DOIUrl":"https://doi.org/10.1109/TETCI.2026.3657946","url":null,"abstract":"Anomaly detection is an imperative but challenging task in intelligent surveillance due to the scarcity and irregularity. Most existing unsupervised methods mainly detect anomalies through frame prediction, but their discriminative distance is limited. Meanwhile, some issues such as lack of semantic correlation and less attention to inter-frame relations are also existed, which are important to anomaly detection. To solve these issues, a multi-perceptual framework (MPFork) using textual and visual information is proposed, which combines the ideas of fusion prediction, semantic perception, and temporal discrimination. Specifically, features fusion prediction module is devised to concentrate on representing the normality and repressing anomalies by fusing heterogeneous features. Meanwhile, an image-text semantic perception module is proposed for semantic consistency by correlating the visual with textual features, where the local and global inferences are adopted to perceive local anomalies, simultaneously measuring the belongingness of each object in global frame. Moreover, a temporal attention discrimination module is developed to capture the inter-frame relations and identify pseudo-abnormal sequences from normal ones in temporal direction. Extensive experiments show the proposed MPFork outshines most of state-of-the-art methods, which indicates the efficacy of our MPFork for anomaly detection.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"10 2","pages":"1821-1833"},"PeriodicalIF":6.0,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148511351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0