Computer Vision - Eccv 2022: 17th European Conference, Tel Aviv, Israel, October 23-27, 2022, Proceedings, Part XXIX

ISBN-13: 9783031198175 / Angielski / Miękka / 2022 / 757 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

​Box-Supervised Instance Segmentation with Level Set Evolution.- Point Primitive Transformer for Long-Term 4D Point Cloud Video Understanding.- Adaptive Agent Transformer for Few-Shot Segmentation.- Waymo Open Dataset: Panoramic Video Panoptic Segmentation.- TransFGU: A Top-down Approach to Fine-Grained Unsupervised Semantic Segmentation.- AdaAfford: Learning to Adapt Manipulation Affordance for 3D Articulated Objects via Few-Shot Interactions.- Cost Aggregation with 4D Convolutional Swin Transformer for Few-Shot Segmentation.- Fine-Grained Egocentric Hand-Object Segmentation: Dataset, Model, and Applications.- Perceptual Artifacts Localization for Inpainting.- 2D Amodal Instance Segmentation Guided by 3D Shape Prior.- Data Efficient 3D Learner via Knowledge Transferred from 2D Model.- Adaptive Spatial-BCE Loss for Weakly Supervised Semantic Segmentation.- Dense Gaussian Processes for Few-Shot Segmentation.- 3D Instances as 1D Kernels.- TransMatting: Enhancing Transparent Objects Matting with Transformers.- MVSalNet:Multi-View Augmentation for RGB-D Salient Object Detection.- k-Means Mask Transformer.- SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustness.- Adversarial Erasing Framework via Triplet with Gated Pyramid Pooling Layer for Weakly Supervised Semantic Segmentation.- Continual Semantic Segmentation via Structure Preserving and Projected Feature Alignment.- Interclass Prototype Relation for Few-Shot Segmentation.- Slim Scissors: Segmenting Thin Object from Synthetic Background.- Abstracting Sketches through Simple Primitives.- Multi-Scale and Cross-Scale Contrastive Learning for Semantic Segmentation.- One-Trimap Video Matting.- D2ADA: Dynamic Density-Aware Active Domain Adaptation for Semantic Segmentation.- Learning Quality-Aware Dynamic Memory for Video Object Segmentation.- Learning Implicit Feature Alignment Function for Semantic Segmentation.- Quantum Motion Segmentation.- Instance As Identity: A Generic Online Paradigm for Video Instance Segmentation.- Laplacian Mesh Transformer: Dual Attention and Topology Aware Network for 3D Mesh Classification and Segmentation.- Geodesic-Former: A Geodesic-Guided Few-Shot 3D Point Cloud Instance Segmenter.- Union-Set Multi-source Model Adaptation for Semantic Segmentation.- Point MixSwap: Attentional Point Cloud Mixing via Swapping Matched Structural Divisions.- BATMAN: Bilateral Attention Transformer in Motion-Appearance Neighboring Space for Video Object Segmentation.- SPSN: Superpixel Prototype Sampling Network for RGB-D Salient Object Detection.- Global Spectral Filter Memory Network for Video Object Segmentation.- Video Instance Segmentation via Multi-Scale Spatio-Temporal Split Attention Transformer.- RankSeg: Adaptive Pixel Classification with Image Category Ranking for Segmentation.- Learning Topological Interactions for Multi-Class Medical Image Segmentation.- Unsupervised Segmentation in Real-World Images via Spelke Object Inference.- A Simple Baseline for Open-Vocabulary Semantic Segmentation with Pre-trained Vision-Language Model.