Research map / Video / Video segmentation and tracking
Learning To Associate Every Segment for Video Panoptic Segmentation
Check 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 |
|---|---|---|---|
| add joint segment-level and pixel-level correspondence objectives | SiamTrack · VIPER | VPQ: 48.4 VPQ vs. 50.2 VPQ | Helped |
| add joint segment-level and pixel-level correspondence objectives | SiamTrack · Cityscapes-VPS val | VPQ: +1.3%VPQ | Helped |
| add joint segment-level and pixel-level correspondence objectives | SiamTrack · Cityscapes-VPS test | VPQ: +1.2%VPQ | Helped |
| add instance-level and semantic class-level contrast together (ablation) | SiamTrack segment correspondence · VIPER | VPQ: 47.9 -> 49.6 | Helped |
| use instance-level contrast without semantic class-level contrast (ablation) | SiamTrack segment correspondence · VIPER | VPQ: 47.9 -> 48.4 | Helped |
| use semantic class-level contrast without instance-level contrast (ablation) | SiamTrack segment correspondence · VIPER | VPQ: 49.6 -> 47.6 | Hurt |
| add tube-matching loss (ablation) | SiamTrack pixel correspondence · VIPER | VPQ: 48.6 -> 49.6 | Helped |
| add tube-matching loss (ablation) | SiamTrack pixel correspondence · VIPER | TC: 79.3 -> 80.8 | Helped |
| add warping loss alongside tube-matching loss (ablation) | SiamTrack pixel correspondence · VIPER | VPQ: 49.6 -> 49.9 | Helped |
| add warping loss alongside tube-matching loss (ablation) | SiamTrack pixel correspondence · VIPER | TC: 80.8 -> 82.2 | Helped |
| train with a temporal gap of 10 rather than 5 (ablation) | SiamTrack · VIPER | VPQ: 49.8 -> 50.2 | Helped |
| train with a temporal gap of 15 rather than 10 (ablation) | SiamTrack · VIPER | VPQ: 50.2 -> 49.9 | Hurt |
Closest papers by meaning
- Video Panoptic Segmentation (CVPR 2020 · similarity 0.90)
- Unifying Training and Inference for Panoptic Segmentation (CVPR 2020 · similarity 0.89)
- Boosting Video Object Segmentation via Space-Time Correspondence Learning (CVPR 2023 · similarity 0.88)
- Seamless Scene Segmentation (CVPR 2019 · similarity 0.88)
- Scene-Centric Unsupervised Panoptic Segmentation (CVPR 2025 · similarity 0.87)
- Task-aligned Part-aware Panoptic Segmentation through Joint Object-Part Representations (CVPR 2024 · 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.