Research map / Video / Video segmentation and tracking
Mask Propagation for Efficient Video Semantic 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 |
|---|---|---|---|
| propagate key-frame masks with query-based flow instead of optical flow (ablation) | MPVSS · VSPW | mIoU: 1.4% | Helped |
| propagate key-frame masks with query-based flow instead of optical flow (ablation) | MPVSS · Cityscapes | mIoU: 1.8% | Helped |
| initialize flow queries with key-frame learned segment queries rather than random queries (ablation) | Query-based flow · VSPW, Swin-T | mIoU: 1.1% | Helped |
| initialize flow queries with key-frame learned segment queries rather than random queries (ablation) | Query-based flow · VSPW, Swin-B | mIoU: 1.5% | Helped |
| refine query-based flow with pixel-wise flow (ablation) | Query-based flow · Cityscapes, Swin-T | mIoU: 1.1% | Helped |
| use query-guided pixel-wise flow instead of optical flow (ablation) | MPVSS · VSPW, Swin-T | mIoU: 0.1% | Helped |
| use query-guided pixel-wise flow instead of optical flow (ablation) | MPVSS · VSPW, Swin-B | mIoU: 0.2% | Helped |
| increase the motion decoder from 1 to 3 stages (ablation) | MPVSS · VSPW, Swin-T | mIoU: 39.4 -> 39.9 | Helped |
| increase the motion decoder from 3 to 4 stages (ablation) | MPVSS · VSPW, Swin-T | mIoU: 39.9 -> 39.6 | Hurt |
| increase key-frame interval from 2 to 10 (ablation) | MPVSS with query-based flow · VSPW, Swin-T | mIoU: 1.1% | Hurt |
| increase key-frame interval from 2 to 10 (ablation) | Optical-flow mask propagation · VSPW, Swin-T | mIoU: 2.63% | Hurt |
| use query-based flow instead of optical flow (ablation) | MPVSS · Cityscapes | mIoU: 0.5% | Helped |
Closest papers by meaning
- Low-Latency Video Semantic Segmentation (CVPR 2018 · similarity 0.90)
- Semantic Video Segmentation by Gated Recurrent Flow Propagation (CVPR 2018 · similarity 0.90)
- VidSeg: Training-free Video Semantic Segmentation based on Diffusion Models (CVPR 2025 · similarity 0.89)
- Bootstrapping Video Semantic Segmentation Model via Distillation-assisted Test-Time Adaptation (CVPR 2026 · similarity 0.89)
- Learning Spatial-Semantic Features for Robust Video Object Segmentation (ICLR 2025 · similarity 0.88)
- $\mathbf{D^{3}S^{2}}$: Diffusion-Guided Dataset Distillation for Semantic Segmentation (NeurIPS 2026 · 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.