Research map / Video / Video segmentation and tracking
Multi-grained Temporal Prototype Learning for Few-shot Video Object Segmentation
Nian Liu, Kepan Nan, Wangbo Zhao, Yuanwei Liu, Xiwen Yao, Salman Khan, Hisham Cholakkal, Rao Muhammad Anwer, Junwei Han, Fahad Shahbaz Khan
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 clip prototype learning to IPMT baseline (ablation) | IPMT · YouTube-VIS | J mean: 57.8 -> 61.8 | Improved |
| add frame prototype learning on top of clip prototype (ablation) | VIPMT · YouTube-VIS | F mean: 58.8 -> 60.3 | Improved |
| add memory prototype learning (ablation) | VIPMT · YouTube-VIS | J mean: 62.5 -> 63.5 | Improved |
| add CCDS cross-category loss (ablation) | VIPMT · YouTube-VIS | mVC7: 62.8 -> 65.7 | Improved |
| use bidirectional clip-frame prototype communication instead of one-way (ablation) | VIPMT · YouTube-VIS | mVC7: 60.1 -> 62.1 | Improved |
| add structural similarity maps to IoUNet for memory selection (ablation) | IoUNet · YouTube-VIS | J mean: 63.0 -> 63.3 | Improved |
| train IoUNet on synthesized image data instead of real or video masks (ablation) | IoUNet · YouTube-VIS | J mean: 63.2 -> 63.3 | Improved |
| vary IoU threshold for memory selection (0.5 to 0.9) (ablation) | IoUNet memory selection · YouTube-VIS | J mean: 63.3 -> 63.5 | No clear effect |
| VIPMT versus IPMT plus semi-supervised VOS propagation | VIPMT · YouTube-VIS | J mean: 59.3 -> 64.2 | Improved |
| VIPMT versus TTI on a harder, longer-video benchmark | VIPMT · MiniVSPW | mVC7: 24.4 -> 42.1 | Improved |
Closest papers by meaning
- Temporal Transductive Inference for Few-Shot Video Object Segmentation (IJCV 2025 · similarity 0.88)
- Delving Deep Into Many-to-Many Attention for Few-Shot Video Object Segmentation (CVPR 2021 · similarity 0.87)
- Prototype Mixture Models for Few-shot Semantic Segmentation (ECCV 2020 · similarity 0.87)
- MOVE: Motion-Guided Few-Shot Video Object Segmentation (ICCV 2025 · similarity 0.87)
- Beyond Pixel and Object: Part Feature as Reference for Few-Shot Video Object Segmentation (AAAI 2025 · similarity 0.87)
- Foreground-Covering Prototype Generation and Matching for SAM-Aided Few-Shot Segmentation (AAAI 2025 · similarity 0.86)
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.
