Pattern recognition : 44th DAGM German Conference, DAGM GCPR 2022, Konstanz, Germany, September 27-30, 2022, proceedings / / edited by Bjö Andres [and five others]

• Pubblicazione: Cham, Switzerland : , : Springer, , [2022]; ©2022
• Descrizione fisica: 1 online resource (607 pages)
• Disciplina: 152.1423
• Soggetto topico: Pattern perception; Pattern recognition systems
• Persona (resp. second.): AndresBjö
• Nota di contenuto: Intro -- Preface -- Organization -- Contents -- Machine Learning Methods -- InvGAN: Invertible GANs -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Architecture -- 3.2 Objective -- 3.3 Dealing with Resolutions Higher Than the Training Resolution -- 4 Experiments -- 4.1 Semantically Consistent Inversion Using InvGAN -- 4.2 Suitability for Image Editing -- 4.3 Tiling to Boost Resolution -- 4.4 Ablation Studies -- 5 Discussion and Future Work -- 6 Conclusion -- References -- Auto-Compressing Subset Pruning for Semantic Image Segmentation -- 1 Introduction -- 2 Related Work -- 3 Auto-Compressing Subset Pruning -- 3.1 Introduction -- 3.2 Details of the Algorithm -- 4 Experiments -- 5 Summary and Outlook -- 6 Author Contributions, Acknowledgements and Disclosure of Funding -- References -- Improving Robustness and Calibration in Ensembles with Diversity Regularization -- 1 Introduction -- 2 Related Work -- 3 Methods and Metrics -- 3.1 Regularizers -- 3.2 Architectures -- 3.3 Metrics -- 4 Experiments and Results -- 4.1 Datasets, Models, and Training -- 4.2 Diversity Regularization Under Dataset Shift -- 4.3 Out-of-Distribution Detection -- 5 Conclusion -- References -- Unsupervised, Semi-supervised and Transfer Learning -- FastSiam: Resource-Efficient Self-supervised Learning on a Single GPU -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Background -- 3.2 Stabilizing the Training Target -- 4 Experiments -- 4.1 Experimental Setup -- 4.2 Main Results -- 4.3 Ablations -- 5 Conclusion -- References -- Reiterative Domain Aware Multi-target Adaptation -- 1 Introduction -- 2 Related Works -- 3 Proposed Method -- 3.1 Transformer-Based Feature Extractor -- 3.2 Training on the Source Dataset -- 3.3 Target Adaptation -- 3.4 Reiterative Target Adaptation -- 3.5 Emphasizing Source Samples -- 4 Experimental Validation -- 4.1 Datasets. ; 4.2 Evaluation Protocol and Settings -- 4.3 Analyses on Office-Home Dataset -- 4.4 Quantitative Comparisons -- 5 Conclusion -- References -- Augmentation Learning for Semi-Supervised Classification -- 1 Introduction -- 2 Related Work -- 2.1 Semi-Supervised Learning -- 2.2 Augmentation Learning -- 3 Method -- 3.1 Augmentation Policy -- 3.2 Optimization Problem -- 4 Experiment -- 4.1 Implementation Details -- 4.2 Datasets -- 4.3 Comparison to State-of-the-Art -- 4.4 Ablation and Analysis -- 5 Conclusion -- References -- Interpretable Machine Learning -- Multi-attribute Open Set Recognition -- 1 Introduction -- 2 Related Works -- 2.1 Open Set Recognition -- 2.2 Compositional Zero-Shot Learning -- 2.3 Shortcut Learning -- 3 Problem Formulation -- 4 Extension from Common OSR Approaches -- 5 Experiments -- 5.1 Experiment Setup -- 5.2 Shortcut Analysis in Multi-attribute OSR -- 5.3 Effects of Shortcut on Confidence Scores -- 5.4 Measuring Explainability for Unknown Samples -- 5.5 Evaluation on a Realistic Dataset -- 6 Conclusion -- References -- A Syntactic Pattern Recognition Based Approach to Online Anomaly Detection and Identification on Electric Motors -- 1 Introduction -- 2 Related Work -- 3 Preliminaries -- 3.1 Syntactic Pattern Recognition -- 3.2 Variable Order Markov Models -- 3.3 Probabilistic Suffix Trees -- 4 Methodology -- 4.1 Lexical Analysis -- 4.2 Model Learning -- 4.3 Unsupervised Condition Detection -- 4.4 Diagnosis -- 4.5 Complexity -- 4.6 Explainability and Traceability -- 5 Experimental Evaluation -- 5.1 CWRU Bearing Dataset -- 5.2 Case Study: Electric Motor Reliability Tests in iRel40 Project -- 6 Conclusion and Future Work -- References -- Sparse Visual Counterfactual Explanations in Image Space -- 1 Introduction -- 2 Visual Counterfactual Explanations (VCEs) -- 2.1 Formulation and Properties of VCEs. ; 2.2 Generation and evaluation of VCEs -- 2.3 Sparse VCEs via the l1.5-metric -- 3 Auto-Frank-Wolfe for lp-VCEs -- 4 Finding Spurious Features with l1.5-VCEs -- 5 Discussion and Limitations -- References -- Low-Level Vision and Computational Photography -- Blind Single Image Super-Resolution via Iterated Shared Prior Learning*-6pt -- 1 Introduction -- 2 Preliminaries -- 2.1 Discrete Wasserstein Distance -- 2.2 Kernel Representation Using Gabor Filters -- 3 Proposed Method -- 3.1 Initial Kernel Estimation -- 3.2 Shared Prior Learning -- 3.3 Supervised Kernel Estimation -- 4 Experimental Results for Blind SISR -- 4.1 SISR with Multiple Bins -- 4.2 One-Shot SISR -- 5 Conclusions -- References -- A Bhattacharyya Coefficient-Based Framework for Noise Model-Aware Random Walker Image Segmentation -- 1 Introduction -- 2 Related Work -- 3 Framework -- 3.1 Random Walker Segmentation -- 3.2 Weight Function -- 3.3 Neighborhood -- 4 Application to Concrete Noise Models -- 4.1 Poisson Noise -- 4.2 Multivariate Gaussian Noise with Constant Covariance -- 4.3 Gaussian Noise with Signal-Dependent Variance -- 5 Experimental Results -- 5.1 Results on Synthetic Data -- 5.2 Results on Real World Data -- 6 Conclusion -- References -- Optimizing Edge Detection for Image Segmentation with Multicut Penalties -- 1 Introduction -- 2 Related Work -- 3 Learning of Edge Weights for Graph Decompositions -- 3.1 Cycle Constraints in the Multicut Problem -- 3.2 Incorporating Cycle Constraints into a CRF -- 3.3 Cooling Mean-Field Updates -- 4 Richer Convolutional Features - CRF -- 5 Experiments -- 5.1 Berkeley Segmentation Dataset and Benchmark -- 5.2 Neuronal Structure Segmentation -- 6 Conclusion -- References -- Hyperspectral Demosaicing of Snapshot Camera Images Using Deep Learning -- 1 Introduction -- 2 Related Work -- 3 Network Architecture -- 4 Data Acquisition. ; 5 Training and Network Evaluation Metrics -- 6 Experiments and Results -- 6.1 Quantitative Results -- 6.2 Qualitative Results -- 6.3 Intraoperative Image Demosaicing -- 7 Discussion and Conclusion -- References -- Motion, Pose Estimation and Tracking -- SF2SE3: Clustering Scene Flow into SE(3)-Motions via Proposal and Selection -- 1 Introduction -- 2 Related Work -- 2.1 Non-rigid Models -- 2.2 Object-Rigid Models -- 3 Approach -- 3.1 Algorithm Outline -- 3.2 Consensus Models -- 3.3 Proposals via Rigidity Constraint -- 3.4 Selection via Coverage Problem -- 3.5 Deduction of Odometry, Image Segmentation, and Scene Flow -- 4 Results -- 5 Discussion -- 6 Conclusion -- References -- Improving Unsupervised Label Propagation for Pose Tracking and Video Object Segmentation -- 1 Introduction -- 2 Related Work -- 3 Label Propagation -- 3.1 General Task -- 3.2 Common Extensions -- 4 Applications and Case Studies -- 4.1 Pose Tracking -- 4.2 Joint Tracking and Keypoint Propagation -- 4.3 Unsupervised Zero-Shot VOS -- 5 Experiments -- 5.1 Ablation Studies -- 5.2 Label Propagation -- 6 Conclusions -- References -- D-InLoc++: Indoor Localization in Dynamic Environments -- 1 Introduction -- 2 Related Work -- 3 Dynamic Object Filtering -- 4 Image Selection Step -- 5 Dynamic Dataset Generation -- 6 Experimental Evaluation -- 6.1 Composition of Test Datasets -- 6.2 Evaluation -- 7 Conclusion -- References -- Global Hierarchical Attention for 3D Point Cloud Analysis -- 1 Introduction -- 2 Related Work -- 3 Global Hierarchical Attention (GHA) -- 3.1 Background: Dot-Product Attention -- 3.2 GHA for Point Cloud Analysis -- 3.3 GHA Block -- 4 Evaluation -- 4.1 Voxel-GHA: 3D Semantic Segmentation -- 4.2 Voxel-GHA: 3D Object Detection -- 4.3 Point-GHA: 3D Object Detection -- 5 Conclusion -- References -- 3D Vision and Stereo. ; InterCap: Joint Markerless 3D Tracking of Humans and Objects in Interaction -- 1 Introduction -- 2 Related Work -- 3 InterCap Method -- 3.1 Multi-Kinect Setup -- 3.2 Sequential Object-Only Tracking -- 3.3 Sequential Human-Only Tracking -- 3.4 Joint Human-Object Tracking over All Frames -- 4 InterCap Dataset -- 5 Experiments -- 6 Discussion -- References -- Online Marker-Free Extrinsic Camera Calibration Using Person Keypoint Detections -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Preprocessing -- 3.2 Data Association -- 3.3 Factor Graph Optimization -- 3.4 Camera Pose Refinement -- 4 Evaluation -- 5 Conclusion -- References -- NeuralMeshing: Differentiable Meshing of Implicit Neural Representations -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Modified Neural Implicit Representation -- 3.2 Iterative Meshing Procedure -- 3.3 Merging Surface Patches -- 3.4 Vertex Prediction -- 4 Evaluation -- 4.1 Reconstruction Quality -- 4.2 Triangle Mesh Properties -- 5 Conclusion -- References -- Detection and Recognition -- Image-Based Detection of Structural Defects Using Hierarchical Multi-scale Attention -- 1 Introduction -- 2 Related Work -- 2.1 Anomaly Detection -- 2.2 Transfer Learning -- 3 Data -- 4 Hierarchical Multi-scale Attention -- 5 Results -- 5.1 Metrics -- 5.2 Scales -- 5.3 Attention -- 5.4 Benchmark -- 6 Conclusion -- References -- A Dataset for Analysing Complex Document Layouts in the Digital Humanities and Its Evaluation with Krippendorff's Alpha -- 1 Introduction -- 2 Related Work -- 3 Evaluation Annotation Consistency -- 3.1 Data Quality Metric -- 3.2 Method Properties -- 4 TexBiG Dataset -- 4.1 Dataset Design -- 4.2 Dataset Construction -- 4.3 Dataset Analysis -- 4.4 Dataset Quality Evaluation -- 4.5 Benchmarking -- 5 Conclusion -- References -- End-to-End Single Shot Detector Using Graph-Based Learnable Duplicate Removal. ; 1 Introduction.
• Titolo autorizzato: Pattern Recognition
• ISBN: 3-031-16788-0
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Lingua di pubblicazione: Inglese
• Record Nr.: 9910595040203321
• Lo trovi qui: Univ. Federico II
• Serie: Lecture Notes in Computer Science