Research map / Video / Video segmentation and tracking
Occluded Video Instance Segmentation: Dataset and ICCV 2021 Challenge
Jiyang Qi, Yan Gao, Yao Hu, Xinggang Wang, Xiaoyu Liu, Xiang Bai, Serge Belongie, Alan Yuille, Philip Torr, Song Bai
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 |
|---|---|---|---|
| Replace random reference-frame sampling with limited sampling from 5 neighboring frames (ablation) | MaskTrack R-CNN · OVIS validation set | AP: 10.8 -> 13.5 | Improved |
| Replace ResNet-50 with Swin-S backbone (ablation) | MaskTrack R-CNN + LSS · OVIS validation set | AP: 13.5 -> 21.1 | Improved |
| Replace ResNet-50 with Swin-S backbone (ablation) | MaskTrack R-CNN + LSS · OVIS validation set | AP_HO: 3.3 -> 6.6 | Improved |
| Add stochastic weight averaging training (ablation) | Ach · OVIS validation set | AP: 28.0 -> 28.9 | Improved |
| Add synthesized image pairs from COCO for training (ablation) | STEm-Seg · OVIS validation set | AP: 13.8 -> 16.2 | Improved |
| Replace backbone with ResNet-152 | STMask · OVIS validation set | AP: 15.4 -> 19.7 | Improved |
| Use Swin-L, larger input resolution and tuned schedule instead of Swin-S (ablation) | MaskTrack R-CNN + LSS · OVIS validation set | AP: 21.1 -> 28.0 | Improved |
| Evaluate on OVIS instead of YouTube-VIS | SipMask · YouTube-VIS vs OVIS test | AP: 32.5 -> 11.7 | Hurt |
| Evaluate on heavily occluded instances instead of all instances | SipMask · OVIS test set | AP: 11.7 -> 3.2 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- STEm-Seg — STEm-Seg: Spatio-temporal Embeddings for Instance Segmentation in Videos (ECCV 2020) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Occluded Video Instance Segmentation: A Benchmark (IJCV 2022 · similarity 0.96)
- LVIS: A Dataset for Large Vocabulary Instance Segmentation (CVPR 2019 · similarity 0.87)
- Robust Instance Segmentation Through Reasoning About Multi-Object Occlusion (CVPR 2021 · similarity 0.86)
- CAVIS: Context-Aware Video Instance Segmentation (ICCV 2025 · similarity 0.86)
- Occlusion-Aware Instance Segmentation Via BiLayer Network Architectures (TPAMI 2023 · similarity 0.86)
- OpenVIS: Open-vocabulary Video Instance Segmentation (AAAI 2025 · similarity 0.85)
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.
