Research map / Video / Video segmentation and tracking
A Generalized Framework for Video Instance Segmentation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use GenVIS online instead of IDOL | ResNet-50 VIS · YouTube-VIS 2021 validation | AP: 43.9 -> 47.1 | Helped |
| Use GenVIS online instead of IDOL | ResNet-50 VIS · OVIS validation | AP: 30.2 -> 35.8 | Helped |
| Use GenVIS online instead of VITA | ResNet-50 VIS · YouTube-VIS 2022 long videos | AP: 32.6 -> 37.5 | Helped |
| Use GenVIS semi-online instead of IDOL | Swin-L VIS · YouTube-VIS 2021 validation | AP: 56.1 -> 60.1 | Helped |
| Use GenVIS semi-online instead of IDOL | Swin-L VIS · OVIS validation | AP: 42.6 -> 45.4 | Helped |
| Use GenVIS semi-online instead of IDOL | Swin-L VIS · YouTube-VIS 2019 validation | AP: 64.3 -> 63.8 | Hurt |
| Add correspondence learning across two clips (ablation) | GenVIS training · OVIS validation | AP: 24.1 -> 27.4 | Helped |
| Add correspondence learning across two clips (ablation) | GenVIS training · YouTube-VIS 2022 long videos | AP: 25.4 -> 29.6 | Helped |
| Train correspondence across five rather than two clips (ablation) | GenVIS training · OVIS validation | AP: 27.4 -> 29.8 | Helped |
| Add UVLA to five-clip correspondence learning (ablation) | GenVIS training · OVIS validation | AP: 29.8 -> 32.5 | Helped |
| Add UVLA to five-clip correspondence learning (ablation) | GenVIS training · YouTube-VIS 2022 long videos | AP: 32.8 -> 35.4 | Helped |
| Add instance prototype memory (ablation) | GenVIS with five-clip UVLA training · OVIS validation | AP: 32.5 -> 33.4 | Helped |
Closest papers by meaning
- MinVIS: A Minimal Video Instance Segmentation Framework without Video-based Training (NeurIPS 2022 · similarity 0.92)
- SyncVIS: Synchronized Video Instance Segmentation (NeurIPS 2024 · similarity 0.91)
- Minimizing Labeled, Maximizing Unlabeled: An Image-Driven Approach for Video Instance Segmentation (CVPR 2025 · similarity 0.91)
- OpenVIS: Open-vocabulary Video Instance Segmentation (AAAI 2025 · similarity 0.87)
- Robust and Consistent Online Video Instance Segmentation via Instance Mask Propagation (AAAI 2025 · similarity 0.87)
- VideoCutLER: Surprisingly Simple Unsupervised Video Instance Segmentation (CVPR 2024 · similarity 0.87)
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.