Research map / Video / Video segmentation and tracking
Semantic Video Segmentation by Gated Recurrent Flow Propagation
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 |
|---|---|---|---|
| Add GRFP temporal propagation to Dilation10 | Dilation10 · CityScapes test | mIoU class: 67.1 -> 67.8 | Helped |
| Add GRFP temporal propagation to LRR-4x | LRR-4x · CityScapes test | mIoU class: 71.8 -> 72.8 | Helped |
| Add GRFP temporal propagation to Dilation10 | Dilation10 · CityScapes validation | mIoU: 0.687 -> 0.694 | Helped |
| Increase inference context from one frame to two frames (ablation) | GRFP · CityScapes validation | mIoU: 0.687 -> 0.691 | Helped |
| Increase inference context from four frames to five frames (ablation) | GRFP · CityScapes validation | mIoU: 0.694 -> 0.694 | No clear effect |
| Add backward propagation with tied forward parameters and T=2 (ablation) | GRFP(5) · CityScapes validation | mIoU: 0.694 -> 0.694 | No clear effect |
| Extend forward-backward propagation from T=2 to T=4 (ablation) | fwbw(a) · CityScapes validation | mIoU: 0.694 -> 0.696 | Helped |
| Refine backward model parameters (ablation) | fwbw(b) · CityScapes validation | mIoU: 0.696 -> 0.695 | Hurt |
| Refine forward, backward, and unary parameters jointly (ablation) | fwbw(c) · CityScapes validation | mIoU: 0.695 -> 0.694 | Hurt |
| Replace the flow source with Flownet (ablation) | GRFP(5) · CityScapes validation | mIoU: 0.694 -> 0.692 | Hurt |
| Jointly train the flow, Dilation10, and recurrent networks (ablation) | GRFP with Flownet · CityScapes validation | mIoU: 0.694 -> 0.690 | Hurt |
| Add GRFP temporal propagation to LRR | LRR · CityScapes validation | mIoU: 0.725 -> 0.736 | Helped |
Closest papers by meaning
- Low-Latency Video Semantic Segmentation (CVPR 2018 · similarity 0.91)
- Every Frame Counts: Joint Learning of Video Segmentation and Optical Flow (AAAI 2020 · similarity 0.90)
- VidSeg: Training-free Video Semantic Segmentation based on Diffusion Models (CVPR 2025 · similarity 0.88)
- Exploring Temporal Feature Correlation for Efficient and Stable Video Semantic Segmentation (AAAI 2024 · similarity 0.87)
- Exploiting Temporal State Space Sharing for Video Semantic Segmentation (CVPR 2025 · similarity 0.87)
- Pay Attention to Target: Relation-Aware Temporal Consistency for Domain Adaptive Video Semantic Segmentation (AAAI 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.