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

ISBN-13: 9783031200762 / Angielski / Miękka / 2022 / 755 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

BEVFormer: Learning Bird’s-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers.- Category-Level 6D Object Pose and Size Estimation Using

Self-Supervised Deep Prior Deformation Networks.- Dense Teacher: Dense Pseudo-Labels for Semi-Supervised Object Detection.- Point-to-Box Network for Accurate Object Detection via Single Point Supervision.- Domain Adaptive Hand Keypoint and Pixel Localization in the Wild.- Towards Data-Efficient Detection Transformers.- Open-Vocabulary DETR with Conditional Matching.- Prediction-Guided Distillation for Dense Object Detection.- Multimodal Object Detection via Probabilistic Ensembling.- Exploiting Unlabeled Data with Vision and Language Models for Object Detection.- CPO: Change Robust Panorama to Point Cloud Localization.- INT: Towards Infinite-Frames 3D Detection with an Efficient Framework.- End-to-End Weakly Supervised Object Detection with Sparse Proposal Evolution.- Calibration-Free Multi-View Crowd Counting.- Unsupervised Domain Adaptation for Monocular 3D Object Detection via Self-Training.- SuperLine3D: Self-Supervised Line Segmentation and Description for LiDAR Point Cloud.- Exploring Plain Vision Transformer Backbones for Object Detection.- Adversarially-Aware Robust Object Detector.- HEAD: HEtero-Assists Distillation for Heterogeneous Object Detectors.- You Should Look at All Objects.- Detecting Twenty-Thousand Classes Using Image-Level Supervision.- DCL-Net: Deep Correspondence Learning Network for 6D Pose Estimation.- Monocular 3D Object Detection with Depth from Motion.- DISP6D: Disentangled Implicit Shape and Pose Learning for Scalable 6D Pose Estimation.- Distilling Object Detectors with Global Knowledge.- Unifying Visual Perception by Dispersible Points Learning.- PseCo: Pseudo Labeling and Consistency Training for Semi-Supervised Object Detection.- Exploring Resolution and Degradation Clues As Self-Supervised Signal for Low Quality Object Detection.- Robust Category-Level 6D Pose Estimation with Coarse-to-Fine Rendering of Neural Features.- Translation, Scale and Rotation: Cross-Modal Alignment Meets RGB-Infrared Vehicle Detection.- RFLA: Gaussian Receptive Field Based Label Assignment for Tiny Object Detection.- Rethinking IoU-Based Optimization for Single-Stage 3D Object Detection.- TD-Road: Top-Down Road Network Extraction with Holistic Graph Construction.- Multi-faceted Distillation of Base-Novel Commonality for Few-Shot Object Detection.- PointCLM: A Contrastive Learning-Based Framework for Multi-Instance Point Cloud Registration.- Weakly Supervised Object Localization via Transformer with Implicit Spatial Calibration.- MTTrans: Cross-Domain Object Detection with Mean Teacher Transformer.- Multi-Domain Multi-Definition Landmark Localization for Small Datasets.- DEVIANT: Depth EquiVarIAnt NeTwork for Monocular 3D Object Detection.- Label-Guided Auxiliary Training Improves 3D Object Detector.- PromptDet: Towards Open-Vocabulary Detection Using Uncurated Images.- Densely Constrained Depth Estimator for Monocular 3D Object Detection.- Polarimetric Pose Prediction.