Research map / Label-efficient and robust learning / Semi-supervised learning
Warp-Refine Propagation: Semi-Supervised Auto-Labeling via Cycle-Consistency
Aditya Ganeshan, Alexis Vallet, Yasunori Kudo, Shin-ichi Maeda, Tommi Kerola, Rares Ambrus, Dennis Park, Adrien Gaidon
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 learned refinement network to warp-inpaint labels (ablation) | warp-inpaint · NYU-V2 | mIoU: 51.28 -> 52.12 | Improved |
| add learned refinement network to warp-inpaint labels (ablation) | warp-inpaint · ApolloScape | mIoU: 72.97 -> 73.74 | Improved |
| add semantic inpainting to motion-only labels (ablation) | motion-only · NYU-V2 | mIoU: 50.38 -> 51.28 | Improved |
| train with warp-refine auto-labels instead of none | MSA-HRNet-OCR · NYU-V2 | mIoU: 50.58 -> 52.12 | Improved |
| train with warp-refine auto-labels instead of none | MSA-HRNet-OCR · ApolloScape | mIoU: 72.63 -> 73.74 | Improved |
| train with warp-refine auto-labels instead of none | MSA-HRNet-OCR · Cityscapes val | mIoU: 83.35 -> 84.07 | Improved |
| train with motion-only auto-labels | MSA-HRNet-OCR · Cityscapes val | mIoU: 83.35 -> 83.01 | Hurt |
| train with motion-only auto-labels | MSA-HRNet-OCR · NYU-V2 | mIoU: 50.58 -> 50.38 | Hurt |
| train with semantic-only auto-labels | MSA-HRNet-OCR · NYU-V2 | mIoU: 50.58 -> 50.93 | Improved |
| use warp-refine instead of semantic-only on a high-performing baseline | OCR Net ResNeXt-50 · Cityscapes val (no extra data) | mIoU: 80.89 -> 80.80 | No clear effect |
| use warp-refine instead of semantic-only | DeepLab V3 ResNeXt-50 · Cityscapes val (no extra data) | mIoU: 80.45 -> 80.68 | Improved |
| use ground-truth labels instead of warp-refine propagated labels (oracle) | MSA-HRNet-OCR · ApolloScape | mIoU: 73.74 -> 75.36 | Improved |
Closest papers by meaning
- Semantic Video Segmentation by Gated Recurrent Flow Propagation (CVPR 2018 · similarity 0.90)
- Improving Semantic Segmentation via Video Propagation and Label Relaxation (CVPR 2019 · similarity 0.89)
- Learning Random-Walk Label Propagation for Weakly-Supervised Semantic Segmentation (CVPR 2017 · similarity 0.88)
- Improving Semi-Supervised and Domain-Adaptive Semantic Segmentation with Self-Supervised Depth Estimation (IJCV 2023 · similarity 0.88)
- Domain Adaptive Video Segmentation via Temporal Pseudo Supervision (ECCV 2022 · similarity 0.88)
- ScaleMatch: Multi-scale Consistency Enhancement for Semi-supervised Semantic Segmentation (AAAI 2025 · similarity 0.85)
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.
