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

ISBN-13: 9783031198113 / Angielski / Miękka / 2022 / 751 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

Relative Contrastive Loss for Unsupervised Representation Learning.- Fine-Grained Fashion Representation Learning by Online Deep Clustering.- NashAE: Disentangling Representations through Adversarial Covariance Minimization.- A Gyrovector Space Approach for Symmetric Positive Semi-Definite Matrix Learning.- Learning Visual Representation from Modality-Shared Contrastive Language-Image Pre-training.- Contrasting Quadratic Assignments for Set-Based Representation Learning.- Class-Incremental Learning with Cross-Space Clustering and Controlled Transfer.- Object Discovery and Representation Networks.- Trading Positional Complexity vs Deepness in Coordinate Networks.- MVDG: A Unified Multi-View Framework for Domain Generalization.- Panoptic Scene Graph Generation.- Object-Compositional Neural Implicit Surfaces.- RigNet: Repetitive Image Guided Network for Depth Completion.- FADE: Fusing the Assets of Decoder and Encoder for Task-Agnostic Upsampling.- LiDAL: Inter-Frame Uncertainty Based Active Learning for 3D LiDAR Semantic Segmentation.- Hierarchical Memory Learning for Fine-Grained Scene Graph Generation.- DODA: Data-Oriented Sim-to-Real Domain Adaptation for 3D Semantic Segmentation.- MTFormer: Multi-task Learning via Transformer and Cross Task Reasoning.- MonoPLFlowNet: Permutohedral Lattice FlowNet for Real-Scale 3D Scene Flow Estimation with Monocular Images.- TO-Scene: A Large-Scale Dataset for Understanding 3D Tabletop Scenes.- Is It Necessary to Transfer Temporal Knowledge for Domain Adaptive Video Semantic Segmentation?.- Meta Spatio-Temporal Debiasing for Video Scene Graph Generation.- Improving the Reliability for Confidence Estimation.- Fine-Grained Scene Graph Generation with Data Transfer.- Pose2Room: Understanding 3D Scenes from Human Activities.- Towards Hard-Positive Query Mining for DETR-Based Human-Object Interaction Detection.- Discovering Human-Object Interaction Concepts via Self-Compositional Learning.- Primitive-Based Shape Abstraction via Nonparametric Bayesian Inference.- Stereo Depth Estimation with Echoes.- Inverted Pyramid Multi-task Transformer for Dense Scene Understanding.- PETR: Position Embedding Transformation for Multi-View 3D Object Detection.- S2Net: Stochastic Sequential Pointcloud Forecasting.- RA-Depth: Resolution Adaptive Self-Supervised Monocular Depth Estimation.- PolyphonicFormer: Unified Query Learning for Depth-Aware Video Panoptic Segmentation.- SQN: Weakly-Supervised Semantic Segmentation of Large-Scale 3D

Point Clouds.- PointMixer: MLP-Mixer for Point Cloud Understanding.- Initialization and Alignment for Adversarial Texture Optimization.- MOTR: End-to-End Multiple-Object Tracking with TRansformer.- GALA: Toward Geometry-and-Lighting-Aware Object Search for Compositing.- LaLaLoc++: Global Floor Plan Comprehension for Layout Localisation in Unvisited Environments.- 3D-PL: Domain Adaptive Depth Estimation with 3D-Aware Pseudo-Labeling.- Panoptic-PartFormer: Learning a Unified Model for Panoptic Part Segmentation.