Research map / Video / Video segmentation and tracking
Self-supervised Amodal Video Object Segmentation
Jian Yao, Yuxin Hong, Chiyu Wang, Tianjun Xiao, Tong He, Francesco Locatello, David Wipf, Yanwei Fu, Zheng Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add test-time adaptation (ablation) | SaVos · FISHBOWL unseen fish, overall | mean-IoU (occluded): 0.5905 -> 0.6886 | Improved |
| add test-time adaptation | Self-Deocclusion · FISHBOWL unseen fish, overall | mean-IoU (occluded): 0.5859 -> 0.5971 | Improved |
| add test-time adaptation (ablation) | SaVos · FISHBOWL unseen fish, overall | mean-IoU (full): 0.8428 -> 0.8663 | Improved |
| add test-time adaptation on unseen BigFishB (ablation) | SaVos · FISHBOWL BigFishB | mean-IoU (occluded): 0.4978 -> 0.6230 | Improved |
| add test-time adaptation on unseen SmallFishA | Self-Deocclusion · FISHBOWL SmallFishA | mean-IoU (occluded): 0.4757 -> 0.4629 | Hurt |
| add test-time adaptation to KINS-Video-Car for FISHBOWL-trained model (ablation) | SaVos · KINS-Video-Car | mean-IoU (occluded): 0.0994 -> 0.2976 | Improved |
| add test-time adaptation to KINS-Video-Car for FISHBOWL-trained model | Self-Deocclusion · KINS-Video-Car | mean-IoU (occluded): 0.0658 -> 0.0969 | Improved |
| replace Self-Deocclusion trained on KINS-Video-Car with SaVos trained on FISHBOWL plus adaptation | Self-Deocclusion · KINS-Video-Car | mean-IoU (occluded): 0.1790 -> 0.2976 | Improved |
| use SaVos instead of Self-Deocclusion | Self-Deocclusion · Chewing Gum | mean-IoU (occluded): 0.3307 -> 0.8046 | Improved |
| use SaVos instead of Self-Deocclusion | Self-Deocclusion · FISHBOWL | mean-IoU (occluded): 0.6502 -> 0.7155 | Improved |
| use SaVos instead of Self-Deocclusion | Self-Deocclusion · KINS-Video-Car | mean-IoU (occluded): 0.1790 -> 0.3132 | Improved |
| use supervised oracle instead of SaVos | SaVos · Chewing Gum | mean-IoU (occluded): 0.8046 vs 0.3321 | No clear effect |
Closest papers by meaning
- Using Diffusion Priors for Video Amodal Segmentation (CVPR 2025 · similarity 0.91)
- BLADE: Box-Level Supervised Amodal Segmentation through Directed Expansion (AAAI 2024 · similarity 0.90)
- Segment Anything, Even Occluded (CVPR 2025 · similarity 0.89)
- Amodal Ground Truth and Completion in the Wild (CVPR 2024 · similarity 0.89)
- Towards Efficient Foundation Model for Zero-shot Amodal Segmentation (CVPR 2025 · similarity 0.89)
- pix2gestalt: Amodal Segmentation by Synthesizing Wholes (CVPR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
