Computer vision - ECCV 2022 . Part XXX : 17th European Conference, Tel Aviv, Israel, October 23-27, 2022 : proceedings / / Shai Avidan [and four others]

• Pubblicazione: Cham, Switzerland : , : Springer, , [2022]; ©2022
• Descrizione fisica: 1 online resource (801 pages)
• Disciplina: 006.4
• Soggetto topico: Pattern recognition systems; Computer vision
• Persona (resp. second.): AvidanShai
• Nota di contenuto: Intro -- Foreword -- Preface -- Organization -- Contents - Part XXX -- Fast Two-View Motion Segmentation Using Christoffel Polynomials -- 1 Introduction -- 2 Related Work -- 2.1 Sampling Based Approaches -- 2.2 Model Fitting Based Approaches -- 3 Notation -- 4 Problem Setup -- 4.1 Two View Motion Segmentation as Algebraic Variety Clustering -- 4.2 Approximating Support Sets via Christoffel Polynomials -- 5 Methodology and Algorithm -- 5.1 One at a Time Algebraic Clustering -- 5.2 Refinements for Two View Motion Segmentation -- 6 Experimental Evaluation -- 6.1 Adelaide-F -- 6.2 Hopkins-Clean (H-C) and Hopkins-Outliers (H-O) -- 6.3 KT3D -- 6.4 BC and BCD -- 6.5 Pairwise -- 7 Conclusions -- References -- UCTNet: Uncertainty-Aware Cross-Modal Transformer Network for Indoor RGB-D Semantic Segmentation -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 Overall Architecture -- 3.2 Depth Uncertainty -- 3.3 From Self-attention to Uncertainty-Aware Self-attention -- 3.4 Fusion Module -- 4 Experiments -- 4.1 Implementation Details -- 4.2 Datasets -- 4.3 Ablation Study -- 4.4 Comparison with State-of-the-Arts -- 4.5 Qualitative Results -- 5 Conclusion -- References -- Bi-directional Contrastive Learning for Domain Adaptive Semantic Segmentation -- 1 Introduction -- 2 Related Work -- 3 Approach -- 3.1 Overview -- 3.2 Bi-directional Contrastive Learning -- 3.3 Dynamic Pseudo Labels -- 3.4 Training Loss -- 4 Experiments -- 4.1 Implementation Details -- 4.2 Results -- 5 Conclusion -- References -- Learning Regional Purity for Instance Segmentation on 3D Point Clouds -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Random Scene Toy Dataset -- 3.2 Network Architecture -- 3.3 Multi-task Learning -- 3.4 Regional Purity Guided Clustering Algorithm -- 4 Experiment -- 4.1 Evaluation on Random Scene Toy Dataset -- 4.2 Evaluation on ScanNet Dataset. ; 4.3 Evaluation on S3DIS Dataset -- 5 Ablation Study -- 6 Conclusion -- References -- Cross-Domain Few-Shot Semantic Segmentation -- 1 Introduction -- 2 Related Work -- 3 Benchmark -- 4 Problem Setting -- 5 Model -- 5.1 Pyramid Anchor-Based Transformation Module -- 5.2 Task-Adaptive Fine-Tuning Inference -- 6 Experiment -- 6.1 Evaluation Setup -- 6.2 Evaluation Metric -- 6.3 Implementation Details -- 6.4 Baseline Performance Analysis -- 6.5 Experimental Results of PATNet -- 6.6 Ablation Study -- 7 Conclusion -- References -- Generative Subgraph Contrast for Self-Supervised Graph Representation Learning -- 1 Introduction -- 2 Related Work -- 2.1 Graph Neural Networks -- 2.2 Graph Contrastive Learning -- 3 Proposed Method -- 3.1 Adaptive Subgraph Generation -- 3.2 OT Distance Based Contrastive Learning -- 4 Experiment -- 4.1 Datasets -- 4.2 Implementation Details -- 4.3 Results -- 4.4 Ablation Studies -- 5 Conclusion -- References -- SdAE: Self-distillated Masked Autoencoder -- 1 Introduction -- 2 Methods -- 2.1 Framework -- 2.2 Discussions on the Teacher Branch Feeding -- 2.3 Distillation Strategy -- 3 Experiments -- 3.1 Fine-Tuning on ImageNet-1 k -- 3.2 Semantic Segmentation -- 3.3 Object Detection -- 4 Ablation Studies -- 4.1 Ablation Studies on Each Component -- 4.2 The EMA Strategy -- 4.3 The Multi-fold Masking Strategy -- 4.4 Evaluation of Teacher and Student Models -- 5 Related Work -- 6 Conclusion -- References -- Demystifying Unsupervised Semantic Correspondence Estimation -- 1 Introduction -- 2 Related Work -- 3 Semantic Correspondence Estimation -- 3.1 Problem Setup -- 3.2 Unsupervised Semantic Correspondence Learning -- 3.3 Unsupervised Asymmetric Correspondence Loss -- 4 Evaluation Protocol -- 4.1 Evaluation Metrics -- 4.2 Evaluation Datasets -- 4.3 Implementation Details -- 5 Experiments. ; 5.1 Impact of Unsupervised Correspondence Objective -- 5.2 Impact of Backbone Model and Pre-training Objective -- 5.3 Impact of Pre-training Dataset -- 5.4 Impact of Finetuning Correspondence Dataset -- 5.5 Detailed Error Analysis -- 5.6 Discussion and Limitations -- 6 Conclusion -- References -- .26em plus .1em minus .1emOpen-Set Semi-Supervised Object Detection -- 1 Introduction -- 2 Related Work -- 3 Revisiting Semi-Supervised Object Detection -- 4 Open-Set Semi-Supervised Object Detection -- 4.1 Semantic Expansion in Open-Set Scenarios -- 4.2 Online OOD Detection -- 4.3 Offline OOD Detection -- 5 Experiments -- 5.1 Experimental Setting and Datasets -- 5.2 Comparison Between Online and Offline OOD Detection -- 5.3 Experiments on Open-set Semi-Supervised Object Detection -- 6 Conclusion -- References -- Vibration-Based Uncertainty Estimation for Learning from Limited Supervision -- 1 Introduction -- 2 Related Work -- 2.1 Semi-supervised Learning -- 2.2 Active Learning -- 2.3 Uncertainty Estimation Approaches -- 3 Approach -- 3.1 Learning from Limited Supervision -- 3.2 Vibration-Based Approach -- 3.3 Theoretical Foundation -- 3.4 Application to Different Scenarios of Learning from Limited Supervision -- 4 Experiments -- 4.1 Datasets and Implementation Details -- 4.2 Main Results -- 4.3 Diagnostic Experiments -- 5 Conclusions -- References -- Concurrent Subsidiary Supervision for Unsupervised Source-Free Domain Adaptation -- 1 Introduction -- 2 Related Work -- 3 Approach -- 3.1 Preliminaries -- 3.2 Theoretical Insights -- 3.3 Sticker Intervention Based Subsidiary Task Design -- 3.4 Training Algorithm Design Under Source-Free Constraints -- 4 Experiments -- 4.1 Experimental Setup -- 4.2 Discussion -- 5 Conclusion -- References -- Weakly Supervised Object Localization Through Inter-class Feature Similarity and Intra-class Appearance Consistency. ; 1 Introduction -- 2 Related Works -- 2.1 Class Activation Maps (CAMs) Based WSOL -- 2.2 Pseudo Label Based WSOL -- 2.3 Attention Based WSOL -- 3 Methodology -- 3.1 Pipeline -- 3.2 Baseline -- 3.3 Inter-class and Intra-class Features Analysis -- 3.4 Inter-class Similarity Feature Loss -- 3.5 Intra-class Appearance Consistency -- 3.6 Class-Agnostic Segmentation Stage -- 4 Experiments -- 4.1 Experimental Setup -- 4.2 Comparison with State-of-the-Arts -- 4.3 Ablation Studies -- 5 Conclusion -- References -- Active Learning Strategies for Weakly-Supervised Object Detection -- 1 Introduction -- 2 Related Work -- 3 Proposed Approach -- 3.1 Problem Statement -- 3.2 Active Learning for Weakly-Supervised Object Detection -- 3.3 BiB: An Active Learning Strategy -- 3.4 Training Detectors with both Weak and Strong Supervision -- 4 Experimental Results -- 4.1 Experimental Setup -- 4.2 Ablation Studies -- 4.3 Comparison of Active Strategies -- 4.4 Comparison to the State of the Art -- 5 Conclusion and Future Work -- References -- mc-BEiT: Multi-choice Discretization for Image BERT Pre-training*-3pt -- 1 Introduction -- 2 Related Works -- 2.1 BERT Pre-training with Masked Language Modeling -- 2.2 Self-supervised Visual Pre-training -- 3 Preliminaries -- 3.1 Image BERT Pre-training with Masked Image Modeling -- 3.2 Masked Image Modeling as Single-Choice Classification -- 4 mc-BEiT -- 4.1 Masked Image Modeling as Multi-choice Classification -- 4.2 Multi-choice Visual Discretization -- 5 Experiments -- 5.1 Pre-training Setup -- 5.2 Image Classification -- 5.3 Object Detection and Instance Segmentation -- 5.4 Semantic Segmentation -- 6 Ablation Study -- 6.1 The Temperature Coefficient -- 6.2 The Semantic Equilibrium Coefficient -- 6.3 Masking Strategy -- 6.4 Tokenizer -- 6.5 Visualization -- 7 Conclusion -- References. ; Bootstrapped Masked Autoencoders for Vision BERT Pretraining -- 1 Introduction -- 2 Related Works -- 3 Approach -- 3.1 Encoder -- 3.2 Feature Injection Module -- 3.3 Regressor -- 3.4 Predictor -- 3.5 Objective Function -- 4 Experiments -- 4.1 Implementations -- 4.2 Analysis of BootMAE -- 4.3 ImageNet Classification Comparison -- 4.4 Downstream Tasks -- 5 Discussion and Conclusion -- References -- Unsupervised Visual Representation Learning by Synchronous Momentum Grouping -- 1 Introduction -- 2 Related Work -- 2.1 Handcraft Pretext Task -- 2.2 Instance Discrimination Method -- 2.3 Group Discrimination Method -- 3 Approach -- 3.1 Group-Level Contrastive Learning -- 3.2 Synchronous Momentum Grouping -- 3.3 Compared with Previous Clustering Methods -- 3.4 Implementation Details -- 4 Experiments -- 4.1 Linear Evaluation -- 4.2 Semi-supervised Fine-Tune Evaluation -- 4.3 Transfer to Other Vision Tasks -- 4.4 Ablation Study -- 5 Conclusion -- References -- Improving Few-Shot Part Segmentation Using Coarse Supervision -- 1 Introduction -- 2 Related Work -- 3 A Joint Model of Labeling Styles -- 3.1 Variational EM for Learning -- 3.2 Coarse Supervision from Keypoints and Figure-Ground Mask -- 4 Benchmarks for Evaluation -- 4.1 Bird Part Segmentation Benchmark -- 4.2 Aircraft Part Segmentation Benchmark -- 5 Part Segmentation Algorithms -- 5.1 Baselines -- 5.2 Details for Our Approach -- 6 Results -- 7 Conclusions -- References -- What to Hide from Your Students: Attention-Guided Masked Image Modeling -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 Preliminaries and Background -- 3.2 AttMask: Attention-Guided Token Masking -- 4 Experiments -- 4.1 Setup -- 4.2 Experimental Analysis -- 4.3 Benchmark -- 4.4 Ablation Study -- 5 Conclusion -- References -- Pointly-Supervised Panoptic Segmentation -- 1 Introduction -- 2 Related Works. ; 2.1 Panoptic Segmentation.
• Titolo autorizzato: Computer Vision – ECCV 2022
• ISBN: 3-031-20056-X
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Lingua di pubblicazione: Inglese
• Record Nr.: 996500067103316
• Lo trovi qui: Univ. di Salerno
• Serie: Lecture notes in computer science.