Research map / Video / Video segmentation and tracking
EPIC-KITCHENS VISOR Benchmark: VIdeo Segmentations and Object Relations
Ahmad Darkhalil, Dandan Shan, Bin Zhu, Jian Ma, Amlan Kar, Richard Ely Locke Higgins, Sanja Fidler, David Fouhey, Dima Damen
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 |
|---|---|---|---|
| fine-tune on VISOR after COCO pretraining (vs COCO pretraining only) (ablation) | STM · VISOR VOS Val | J&F: 56.9 -> 75.8 | Improved |
| add COCO pretraining to VISOR-only training (ablation) | STM · VISOR VOS Val | J&F: 62.8 -> 75.8 | Improved |
| fine-tune on VISOR instead of DAVIS+YouTube-VOS after COCO pretraining (ablation) | STM · VISOR VOS Val | J&F: 60.9 -> 75.8 | Improved |
| fine-tune on DAVIS+YouTube-VOS after COCO pretraining (ablation) | STM · VISOR VOS Val | J&F: 56.9 -> 60.9 | Improved |
| evaluate on unseen kitchens vs all kitchens | STM best model · VISOR VOS Test | J&F: 78.0 -> 76.6 | Hurt |
| provide oracle evidence frame to PointRend baseline (ablation) | PointRend WDTCF baseline · VISOR WDTCF | Source accuracy: 28.8 -> 88.3 | Improved |
| provide oracle source prior and query class (ablation) | PointRend WDTCF baseline · VISOR WDTCF | Source accuracy: 28.8 -> 33.8 | Improved |
| add oracle prior to oracle evidence frame (ablation) | PointRend WDTCF baseline · VISOR WDTCF | Source accuracy: 88.3 -> 94.1 | Improved |
Closest papers by meaning
- EPIC Fields: Marrying 3D Geometry and Video Understanding (NeurIPS 2023 · similarity 0.88)
- HD-EPIC: A Highly-Detailed Egocentric Video Dataset (CVPR 2025 · similarity 0.88)
- The EPIC-KITCHENS Dataset: Collection, Challenges and Baselines (TPAMI 2020 · similarity 0.87)
- Learning to Segment Referred Objects from Narrated Egocentric Videos (CVPR 2024 · similarity 0.85)
- MMR: A Large-scale Benchmark Dataset for Multi-target and Multi-granularity Reasoning Segmentation (ICLR 2025 · similarity 0.84)
- OW-VISCapTor: Abstractors for Open-World Video Instance Segmentation and Captioning (NeurIPS 2024 · similarity 0.84)
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.
