Computer Vision – ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XXIV

ISBN-13: 9783031200526 / Angielski / Miękka / 2022 / 747 str.

Shai Avidan; Gabriel Brostow; Moustapha Cissé

Opis

The 39-volume set, comprising the LNCS books 13661 until 13699, constitutes the refereed proceedings of the 17th European Conference on Computer Vision, ECCV 2022, held in Tel Aviv, Israel, during October 23–27, 2022. The 1645 papers presented in these proceedings were carefully reviewed and selected from a total of 5804 submissions. The papers deal with topics such as computer vision; machine learning; deep neural networks; reinforcement learning; object recognition; image classification; image processing; object detection; semantic segmentation; human pose estimation; 3d reconstruction; stereo vision; computational photography; neural networks; image coding; image reconstruction; object recognition; motion estimation.

Spis treści

Improving Vision Transformers by Revisiting High-Frequency Components.- Recurrent Bilinear Optimization for Binary Neural Networks.- Neural Architecture Search for Spiking Neural Networks.- Where to Focus: Investigating Hierarchical Attention Relationship for Fine-Grained Visual Classification.- DaViT: Dual Attention Vision Transformers.- Optimal Transport for Label-Efficient Visible-Infrared Person Re-identification.- Locality Guidance for Improving Vision Transformers on Tiny Datasets.- Neighborhood Collective Estimation for Noisy Label Identification and Correction.- Few-Shot Class-Incremental Learning via Entropy-Regularized Data-Free Replay.- Anti-Retroactive Interference for Lifelong Learning.- Towards Calibrated Hyper-Sphere Representation via Distribution Overlap Coefficient for Long-Tailed Learning.- Dynamic Metric Learning with Cross-Level Concept Distillation.- MENet: A Memory-Based Network with Dual-Branch for Efficient Event Stream Processing.- Out-of-Distribution Detection with Boundary Aware Learning.- Learning Hierarchy Aware Features for Reducing Mistake Severity.- Learning to Detect Every Thing in an Open World.- KVT: k-NN Attention for Boosting Vision Transformers.- Registration Based Few-Shot Anomaly Detection.- Improving Robustness by Enhancing Weak Subnets.- Learning Invariant Visual Representations for Compositional Zero-Shot Learning.- Improving Covariance Conditioning of the SVD Meta-Layer by Orthogonality.- Out-of-Distribution Detection with Semantic Mismatch under Masking.- Data-Free Neural Architecture Search via Recursive Label Calibration.- Learning from Multiple Annotator Noisy Labels via Sample-Wise Label Fusion.- Acknowledging the Unknown for Multi-Label Learning with Single Positive Labels.- AutoMix: Unveiling the Power of Mixup for Stronger Classifiers.- MaxViT: Multi-axis Vision Transformer.- ScalableViT: Rethinking the Context-Oriented Generalization of Vision Transformer.- Three Things Everyone Should Know about Vision Transformers.- DeiT III: Revenge of the ViT.- MixSKD: Self-Knowledge Distillation from Mixup for Image Recognition.- Self-Feature Distillation with Uncertainty Modeling for Degraded Image Recognition.- Novel Class Discovery without Forgetting.- SAFA: Sample-Adaptive Feature Augmentation for Long-Tailed Image Classification.- Negative Samples Are at Large: Leveraging Hard-Distance Elastic Loss for Re-identification.- Discrete-Constrained Regression for Local Counting Models.- Breadcrumbs: Adversarial Class-Balanced Sampling for Long-Tailed Recognition.- Chairs Can Be Stood On: Overcoming Object Bias in Human-Object Interaction Detection.- A Fast Knowledge Distillation Framework for Visual Recognition.- DICE: Leveraging Sparsification for Out-of-Distribution Detection.- Invariant Feature Learning for Generalized Long-Tailed Classification.- Sliced Recursive Transformer.