OneVOS: Unifying Video Object Segmentation with All-in-One Transformer Framework
OneVOS proposes a unified framework for Video Object Segmentation using an All-in-One Transformer, achieving state-of-the-art performance.
Self-supervised Video Object Segmentation with Distillation Learning of Deformable Attention
Proposing a method for self-supervised video object segmentation using distillation learning of deformable attention to address challenges in VOS.
ClickVOS: Click Video Object Segmentation
提案されたClickVOSアプローチは、1-2秒の対象物を示すための単一クリックでビデオ内の対象物をセグメント化する革新的な方法です。
OneVOS: Unifying Video Object Segmentation with All-in-One Transformer Framework
OneVOS proposes a unified framework for Video Object Segmentation using an All-in-One Transformer, achieving state-of-the-art performance across various datasets.
Efficient Video Object Segmentation via Modulated Cross-Attention Memory: A Detailed Analysis
Efficient Video Object Segmentation through Modulated Cross-Attention Memory.
ClickVOS: Click Video Object Segmentation
ClickVOS simplifies video object segmentation with a single click per object, offering practical applications and research implications.
Robust Video Object Segmentation with Object-Level Memory Reading
Cutie, a video object segmentation network, uses object-level memory reading to effectively integrate high-level object representations with low-level pixel features, enabling robust performance in challenging scenarios with occlusions and distractors.
Large-scale Complex and Long Video Object Segmentation Challenge: Methods and Results
The 6th Large-scale Video Object Segmentation (LSVOS) challenge introduced more challenging datasets, MOSE, LVOS, and MeViS, to evaluate the performance of video object segmentation models in complex real-world scenarios. The challenge attracted significant international participation, with 129 teams from over 20 institutes across 8 countries. The top-performing solutions from the VOS and RVOS tracks showcased novel methodologies that leverage memory networks, language models, and spatio-temporal refinement to address the challenges posed by the new datasets.
Efficient Video Object Segmentation Using Extremely Small Foundation Models with Visual Prompting
A simple prompting module can adapt foundational deep learning models to specific video object segmentation tasks, maintaining accuracy and increasing generalization capability on generic deep features.
Jointly Improving Memory Matching and Decoding to Enhance Video Object Segmentation
A novel approach that jointly improves the memory matching and decoding stages to alleviate the false matching issue in video object segmentation.
