LEADER 12080nam 22006495 450 001 996635665503316 005 20250626164116.0 010 $a3-031-78189-9 024 7 $a10.1007/978-3-031-78189-6 035 $a(CKB)36959450600041 035 $a(MiAaPQ)EBC31826293 035 $a(Au-PeEL)EBL31826293 035 $a(DE-He213)978-3-031-78189-6 035 $a(OCoLC)1478697509 035 $a(EXLCZ)9936959450600041 100 $a20241210d2025 u| 0 101 0 $aeng 135 $aur||||||||||| 181 $ctxt$2rdacontent 182 $cc$2rdamedia 183 $acr$2rdacarrier 200 10$aPattern Recognition $e27th International Conference, ICPR 2024, Kolkata, India, December 1?5, 2024, Proceedings, Part IX /$fedited by Apostolos Antonacopoulos, Subhasis Chaudhuri, Rama Chellappa, Cheng-Lin Liu, Saumik Bhattacharya, Umapada Pal 205 $a1st ed. 2025. 210 1$aCham :$cSpringer Nature Switzerland :$cImprint: Springer,$d2025. 215 $a1 online resource (509 pages) 225 1 $aLecture Notes in Computer Science,$x1611-3349 ;$v15309 311 08$a3-031-78188-0 327 $aIntro -- President's Address -- Preface -- Organization -- Contents - Part IX -- Mask and Compress: Efficient Skeleton-Based Action Recognition in Continual Learning -- 1 Introduction -- 2 Related Works -- 3 Method -- 3.1 Preliminaries -- 3.2 CHARON -- 4 Experimental Analysis -- 4.1 Datasets -- 4.2 Implementation Details -- 4.3 Results -- 4.4 Ablations -- 5 Conclusions -- References -- Text-Driven Prototype Learning for Few-Shot Class-Incremental Learning -- 1 Introduction -- 2 Related Works -- 3 Method -- 4 CIFAR100-Text and MiniImageNet-Text Datasets -- 5 Experiments -- 5.1 Datasets and Settings -- 5.2 Text-Driven Prototype Analysis -- 6 Discussions -- 7 Conclusion -- References -- Dual Supervised Contrastive Learning Based on Perturbation Uncertainty for Online Class Incremental Learning -- 1 Introduction -- 2 Related Work -- 2.1 Continual Learning -- 2.2 Contrastive Learning -- 3 Method -- 3.1 Problem Definition -- 3.2 Method Overview -- 3.3 Perturbation Uncertainty Based Memory Retrieval -- 3.4 Supervised Contrastive Learning -- 4 Experiment -- 4.1 Experiment Setup -- 4.2 Performance Comparison -- 4.3 Ablation Study -- 5 Conclusion -- References -- Breaking Information Silos: Global Guided Task Prediction for Class-Incremental Learning -- 1 Introduction -- 2 Related Work -- 3 Methods -- 3.1 Problem Setting and Method Overview -- 3.2 Local De-Redundant Module -- 3.3 Global Information Module -- 3.4 Attention Module -- 3.5 Optimizing and Lightweight Model -- 4 Experiments -- 4.1 Experiment Setup and Implementation Details -- 4.2 Results and Discussion -- 4.3 Ablation Study and Analysis -- 5 Conclusion -- References -- Conditioned Prompt-Optimization for Continual Deepfake Detection -- 1 Introduction -- 2 Related Work -- 3 Preliminaries -- 3.1 Problem Formulation -- 3.2 Prompt Tuning -- 4 Prompt2Guard -- 4.1 Text-Prompt Conditioning. 327 $a4.2 Continual Read-Only Prompts -- 4.3 Predictions Ensembling -- 5 Experiments -- 5.1 Comparative Results -- 5.2 Ablations -- 6 Conclusions -- References -- Plasticity Driven Knowledge Transfer for Continual Deep Reinforcement Learning in Financial Trading -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 Continual Learning Approach -- 3.2 Temporal Focused Sampling Experience Replay -- 3.3 Knowledge Transfer Methodology -- 4 Experimental Evaluation -- 4.1 Dataset and Feature Extraction -- 4.2 Model Architecture -- 4.3 Results -- 5 Conclusions -- References -- Orthogonal Latent Compression for Streaming Anomaly Detection in Industrial Vision -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Overview -- 3.2 Orthogonal Latent Compression -- 3.3 Loss Function -- 3.4 Abnormal Scores -- 4 Experiments -- 4.1 Experimental Setup -- 4.2 Evaluation on Streaming AD -- 4.3 Evaluation on Offline AD -- 4.4 Ablation Analysis -- 5 Conclusion -- References -- Out-of-Distribution Forgetting: Vulnerability of Continual Learning to Intra-class Distribution Shift -- 1 Introduction -- 2 Related Works -- 2.1 Continual Learning -- 2.2 Security Concerns of Neural Networks -- 2.3 Several Concerns of Continual Learning -- 3 Out-of-Distribution Forgetting -- 3.1 Standard CL Paradigm -- 3.2 OODF Paradigm -- 3.3 Introducing of the Distribution Shift -- 4 Experiment Settings -- 5 Properties of OODF -- 5.1 Delayed Effect -- 5.2 Targeting -- 5.3 Continual Detrimental -- 6 Analysis -- 6.1 Occlusion Strength -- 6.2 Various Conditions of Shift -- 6.3 Shift Position in the Learning Sequence -- 6.4 Different Percentage r and Strength -- 6.5 Mechanism of OODF -- 6.6 Proposal for Improving OODF -- 7 Conclusion -- References -- Generating Multi-objective Fronts from Streamed Data Using Nested List -- 1 Introduction -- 1.1 Prerequisites -- 1.2 Literature Survey. 327 $a1.3 Gap Identification and Motivation -- 1.4 Salient Points of the Proposed Approach -- 1.5 Organization of the Paper -- 2 A Nested List Structure for Non-dominated Sorting of Streamed Data Elements -- 2.1 Benchmark Data-Set Specification -- 3 Result and Discussion -- 3.1 Complexity Analysis -- 3.2 Correctness and Completeness -- 4 Conclusion -- References -- Mapping the Unknown: A New Approach to Open-World Video Recognition -- 1 Introduction -- 2 Related Work -- 3 Dynamic Ensembles for OWR -- 3.1 Ensemble Decision Module -- 3.2 Generation Ensemble Module -- 3.3 Update and Limit Module -- 4 Experiments Preliminary -- 4.1 Experiment Dataset Configuration -- 4.2 Experimental Setup -- 5 Experiment Results -- 5.1 Comparison Against State-of-the-Art Face Recognition in OWR -- 5.2 Sensitivity About Parameters -- 6 Conclusions -- References -- ESL: Explain to Improve Streaming Learning for Transformers -- 1 Introduction -- 2 Related Work -- 3 Proposed Framework: ESL -- 3.1 XAI Method: Rollout Feature Explanation Method (RFEM) -- 3.2 Input Patch Selection -- 3.3 Streaming Learner: Entropy-Based Move-To-Data (EMTD) and Retargeting (EMTDR) -- 4 Experiments and Results -- 4.1 Experimental Details -- 4.2 Results -- 4.3 Ablation Study -- 5 Conclusion and Future Work -- References -- Detection of Unknown Errors in Human-Centered Systems -- 1 Introduction -- 1.1 Contributions -- 1.2 Paper Organization -- 2 Preliminaries -- 2.1 Signal Temporal Logic -- 2.2 Physics-Driven Surrogate Model -- 3 Coefficient Mining from Trajectory -- 3.1 Dynamics Induced RNN -- 3.2 Forward Pass in DiH-RNN -- 3.3 Backpropagation to Learn Coefficients -- 4 Conformal Inference -- 5 Case Studies -- 5.1 Automated Insulin Delivery System Example -- 5.2 Aircraft Example -- 5.3 Autonomous Driving Example -- 6 Evaluation Method and Metrics -- 6.1 Unknown-Unknown Scenario Simulation. 327 $a6.2 Baseline Strategy -- 7 Results -- 7.1 Automated Insulin Delivery System Example -- 7.2 Aircraft Example -- 7.3 Autonomous Driving Example -- 8 Future Works -- 9 Conclusions -- References -- Source-Free Test-Time Adaptation For Online Surface-Defect Detection -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Supervisor -- 3.2 Augmented Mean Prediction -- 3.3 Dynamically-Balancing Loss -- 3.4 Model Update Pipeline -- 4 Experiments -- 4.1 Datasets and Pre-training -- 4.2 Results -- 4.3 Ablation Study -- 5 Conclusion -- References -- Alleviating Catastrophic Forgetting in Facial Expression Recognition with Emotion-Centered Models -- 1 Introduction -- 2 Related Works -- 3 Proposed Method -- 3.1 Emotion-Centered Generative Replay -- 3.2 General Pipeline -- 4 Experiments -- 4.1 Results -- 5 Discussion -- 6 Conclusion -- A Appendix -- A.1 Evaluation of the MNIST Dataset -- References -- Satellite State Prediction and Maneuver Detection Analysis Using NCDEs -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 Dataset -- 3.2 Data Preprocessing Step -- 3.3 SSPMDA Architecture -- 4 Empirical Evaluations -- 4.1 Experimental Results -- 4.2 Ablation Study -- 5 Visualization -- 6 Conclusion -- References -- MIXAD: Memory-Induced Explainable Time Series Anomaly Detection -- 1 Introduction -- 2 Related Works -- 3 Method -- 3.1 Problem Formulation -- 3.2 Spatiotemporal Recurrent Convolution Unit -- 3.3 Memory-Augmented Graph Structure Learning -- 3.4 Memory-Induced Explainable Anomaly Detection (MIXAD) -- 3.5 Anomaly Scoring -- 3.6 Anomaly Interpretation -- 4 Experiments and Analysis -- 4.1 Datasets and Baselines -- 4.2 Evaluation Metrics -- 4.3 Performance Comparisons -- 4.4 Ablation Study -- 4.5 Visualization of Node Embeddings -- 4.6 Visualization of Anomaly Scores -- 5 Case Study: Exathlon Dataset and Testbed -- 6 Conclusion -- References. 327 $aRough Set Theoretic Approach for Solving the Multi-Armed Bandit Problems -- 1 Introduction -- 2 Applying Rough Set Concepts to Stochastic Multi-Armed Bandits -- 3 Proposed Methodology -- 4 Experimental Results -- 4.1 Bandit Problem -- 4.2 Advertising Problem -- 4.3 Election Campaign Problem -- 4.4 Ablation Study -- 4.5 Discussions -- 5 Conclusions and Future Work -- References -- Hybrid Graph Representation Learning: Integrating Euclidean and Hyperbolic Space -- 1 Introduction -- 2 Related Work -- 2.1 Representation Learning -- 2.2 Hyperbolic Representation Learning -- 3 Preliminaries -- 3.1 Poincaré Ball Model -- 3.2 Gyrovector Spaces -- 4 Method -- 4.1 Hyperbolic Encoder -- 4.2 Euclidean Loss -- 4.3 Hyperbolic Hierarchy Loss -- 4.4 Hyperbolic Uniformity Loss -- 4.5 Total Loss -- 5 Experiments -- 5.1 Experimental Setup -- 5.2 Node Classification and Clustering -- 5.3 Visualization -- 5.4 Ablation Study -- 6 Conclusion -- References -- Learning Object Focused Attention -- 1 Introduction -- 2 Object Focused Attention -- 3 Self-supervised Option with MAE -- 4 Adjacency Regularization -- 5 Related Work -- 5.1 Transformers and Self-Attention -- 5.2 Holistic Shape Representation -- 5.3 Multi-label Classification -- 6 Experimental Evaluation -- 6.1 Multi-label Classification on MS-COCO and Pascal Voc2012 -- 6.2 Out-of-Distribution Background Corruption with Stable Diffusion -- 6.3 Learning Shape Representations over Textures -- 7 Discussion and Future Work -- References -- Stereographic Projection for Embedding Hierarchical Structures in Hyperbolic Space -- 1 Introduction -- 2 Background -- 2.1 Hyperbolic Neural Networks -- 2.2 Topic Model -- 3 Stereographic Projection Transition Mapping -- 3.1 Limitations of Exponential Mapping for Hierarchical Embeddings -- 3.2 Method: Stereographic Projection Transition Mapping -- 3.3 Optimization Algorithm for SPTM. 327 $a4 Experiment. 330 $aThe multi-volume set of LNCS books with volume numbers 15301-15333 constitutes the refereed proceedings of the 27th International Conference on Pattern Recognition, ICPR 2024, held in Kolkata, India, during December 1?5, 2024. The 963 papers presented in these proceedings were carefully reviewed and selected from a total of 2106 submissions. They deal with topics such as Pattern Recognition; Artificial Intelligence; Machine Learning; Computer Vision; Robot Vision; Machine Vision; Image Processing; Speech Processing; Signal Processing; Video Processing; Biometrics; Human-Computer Interaction (HCI); Document Analysis; Document Recognition; Biomedical Imaging; Bioinformatics. 410 0$aLecture Notes in Computer Science,$x1611-3349 ;$v15309 606 $aComputer vision 606 $aMachine learning 606 $aComputer Vision 606 $aMachine Learning 615 0$aComputer vision. 615 0$aMachine learning. 615 14$aComputer Vision. 615 24$aMachine Learning. 676 $a006.37 700 $aAntonacopoulos$b Apostolos$0885419 701 $aChaudhuri$b Subhasis$0846530 701 $aChellappa$b Rama$0491442 701 $aLiu$b Cheng-Lin$0861045 701 $aBhattacharya$b Saumik$01782600 701 $aPal$b Umapada$01782601 801 0$bMiAaPQ 801 1$bMiAaPQ 801 2$bMiAaPQ 906 $aBOOK 912 $a996635665503316 996 $aPattern Recognition$94309011 997 $aUNISA