Computer Vision – ECCV 2024

Robust Nearest Neighbors for Source-Free Domain Adaptation under Class Distribution Shift.- Chains of Diffusion Models.- Time-Efficient and Identity-Consistent Virtual Try-On Using A Variant of Altered Diffusion Models.- Feature Diversification and Adaptation for Federated Domain Generalization.- Grounding Image Matching in 3D with MASt3R.- TP2O: Creative Text Pair-to-Object Generation using Balance Swap-Sampling.- RoDUS: Robust Decomposition of Static and Dynamic Elements in Urban Scenes.- RecurrentBEV: A Long-term Temporal Fusion Framework for Multi-view 3D Detection.- Efficient Bias Mitigation Without Privileged Information.- MC-PanDA: Mask Confidence for Panoptic Domain Adaptation.- Learning Neural Deformation Representation for 4D Dynamic Shape Generation.- Dynamic Guidance Adversarial Distillation with Enhanced Teacher Knowledge.- Decomposition Betters Tracking Everything Everywhere.- Straightforward Layer-wise Pruning for More Efficient Visual Adaptation.- Synchronization is All You Need: Exocentric-to-Egocentric Transfer for Temporal Action Segmentation with Unlabeled Synchronized Video Pairs.- LAPT: Label-driven Automated Prompt Tuning for OOD Detection with Vision-Language Models.- Domain Shifting: A Generalized Solution for Heterogeneous Cross-Modality Person Re-Identification.- Self-Supervised Video Desmoking for Laparoscopic Surgery.- Removing Rows and Columns of Tokens in Vision Transformer enables Faster Dense Prediction without Retraining.- Continuity Preserving Online CenterLine Graph Learning.- Decomposition of Neural Discrete Representations for Large-Scale 3D Mapping.- MirrorGaussian: Reflecting 3D Gaussians for Reconstructing Mirror Reflections.- Leveraging Representations from Intermediate Encoder-blocks for Synthetic Image Detection.- Exploring Vulnerabilities in Spiking Neural Networks: Direct Adversarial Attacks on Raw Event Data.- HSR: Holistic 3D Human-Scene Reconstruction from Monocular Videos.- Online Video Quality Enhancement with Spatial-Temporal Look-up Tables.- PARIS3D: Reasoning-based 3D Part Segmentation Using Large Multimodal Model.