Research map / Video / Video segmentation and tracking
Domain Adaptive Video Segmentation via Temporal Pseudo Supervision
Yun Xing, Dayan Guan, Jiaxing Huang, Shijian Lu
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 |
|---|---|---|---|
| use cross-frame augmentation and pseudo labelling (TPS) instead of spatial augmentation (PixMatch) | PixMatch · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 51.0 -> 53.8 | Improved |
| use cross-frame augmentation and pseudo labelling (TPS) instead of spatial augmentation (PixMatch) | PixMatch · VIPER to Cityscapes-Seq | mIoU: 46.7 -> 48.9 | Improved |
| use consistency training (TPS) instead of adversarial learning (DA-VSN) | DA-VSN · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 49.5 -> 53.8 | Improved |
| use propagation interval 1 instead of 3 (ablation) | TPS · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 51.7 -> 53.8 | Improved |
| use propagation interval 1 instead of 2 (ablation) | TPS · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 52.3 -> 53.8 | Improved |
| use confidence threshold 0 instead of 0.50 (ablation) | TPS · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 53.1 -> 53.8 | Improved |
| use confidence threshold 0 instead of 0.25 (ablation) | TPS · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 53.5 -> 53.8 | Improved |
| raise target loss weight lambda_T from 0.1 to 1.0 (ablation) | TPS · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 50.0 -> 53.8 | Improved |
| raise target loss weight lambda_T from 1.0 to 2.0 (ablation) | TPS · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 53.8 -> 53.3 | Hurt |
| add TPS consistency training to DA-VSN (ablation) | DA-VSN · SYNTHIA-Seq to Cityscapes-Seq | mIoU: 49.5 -> 55.1 | Improved |
| add TPS consistency training to DA-VSN (ablation) | DA-VSN · VIPER to Cityscapes-Seq | mIoU: 47.8 -> 50.2 | Improved |
Closest papers by meaning
- Pay Attention to Target: Relation-Aware Temporal Consistency for Domain Adaptive Video Semantic Segmentation (AAAI 2024 · similarity 0.93)
- Domain Adaptive Video Segmentation via Temporal Consistency Regularization (ICCV 2021 · similarity 0.90)
- Is It Necessary to Transfer Temporal Knowledge for Domain Adaptive Video Semantic Segmentation? (ECCV 2022 · similarity 0.90)
- Scene Consistency Representation Learning for Video Scene Segmentation (CVPR 2022 · similarity 0.90)
- On the Road to Online Adaptation for Semantic Image Segmentation (CVPR 2022 · similarity 0.89)
- Exploring Temporal Feature Correlation for Efficient and Stable Video Semantic Segmentation (AAAI 2024 · similarity 0.89)
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.
