Research map / Label-efficient and robust learning / Domain adaptation and generalization
Spatio-Temporal Pixel-Level Contrastive Learning-Based Source-Free Domain Adaptation for 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 |
|---|---|---|---|
| Use STPL instead of the strongest listed baseline, TPS | Video SFDA on ACCEL · VIPER → Cityscapes-Seq | mIoU: 48.9 -> 52.5 | Helped |
| Use STPL instead of the strongest listed baseline, TPS | Video adaptation on ACCEL · SYNTHIA-Seq → Cityscapes-Seq | mIoU: 53.8 -> 51.8 | Hurt |
| Replace pseudo-label cross-entropy self-training with pixel-level contrastive learning (ablation) | STPL · VIPER → Cityscapes-Seq | mIoU: 45.4 -> 45.7 | Helped |
| Add temporal-only contrast to duplicate-frame contrast (ablation) | STPL · VIPER → Cityscapes-Seq | mIoU: 45.7 -> 47.4 | Helped |
| Add spatial-only contrast to duplicate-frame contrast (ablation) | STPL · VIPER → Cityscapes-Seq | mIoU: 45.7 -> 51.1 | Helped |
| Combine spatial-only and temporal-only contrastive losses (ablation) | STPL · VIPER → Cityscapes-Seq | mIoU: 51.1 -> 51.4 | Helped |
| Replace the separate spatial and temporal losses with unified spatio-temporal contrast (ablation) | STPL · VIPER → Cityscapes-Seq | mIoU: 51.4 -> 52.5 | Helped |
| Add STAM rather than using concatenation for feature fusion (ablation) | STPL · VIPER → Cityscapes-Seq | mIoU: 52.3 -> 52.5 | Helped |
| Use STPL instead of source-only training | ACCEL · VIPER → Cityscapes-Seq | temporal consistency: 72.93 -> 82.14 | Helped |
Closest papers by meaning
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.89)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.89)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (CVPR 2025 · similarity 0.88)
- Towards Unsupervised Domain Bridging via Image Degradation in Semantic Segmentation (NeurIPS 2025 · similarity 0.88)
- Pay Attention to Target: Relation-Aware Temporal Consistency for Domain Adaptive Video Semantic Segmentation (AAAI 2024 · similarity 0.88)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · 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.