Research map / Label-efficient and robust learning / Domain adaptation and generalization
A Curriculum-Style Self-Training Approach for Source-Free Semantic Segmentation
Yuxi Wang, Jian Liang, Zhaoxiang Zhang
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 complementary curriculum self-training on top of curriculum feature alignment, then propagation (ablation) | ATP (ResNet-101) · GTA5->Cityscapes (source-free) | mIoU: 45.0 -> 49.8 -> 52.6 | Improved |
| add complementary self-training and propagation to alignment (ablation) | ATP (P2T-Base) · GTA5->Cityscapes (source-free) | mIoU: 48.3 -> 52.8 -> 55.8 | Improved |
| add complementary self-training and propagation to alignment (ablation) | ATP (MiT-B5) · GTA5->Cityscapes (source-free) | mIoU: 54.0 -> 62.4 -> 64.0 | Improved |
| add information propagation (easy/hard split semi-supervised post-processing) (ablation) | ATP (ResNet-101) · GTA5->Cityscapes (source-free) | mIoU: 49.8 -> 52.6 | Improved |
| add information propagation (ablation) | ATP (MiT-B5) · GTA5->Cityscapes (source-free) | mIoU: 62.4 -> 64.0 | Improved |
| add complementary self-training and propagation to alignment (ablation) | ATP (ResNet-101) · SYNTHIA->Cityscapes (13-class mIoU*, source-free) | mIoU: 49.0 -> 55.9 -> 57.9 | Improved |
| add information propagation (ablation) | ATP (MiT-B5) · SYNTHIA->Cityscapes (13-class mIoU*, source-free) | mIoU: 64.9 -> 66.6 | Improved |
| add complementary self-training and propagation to alignment (ablation) | ATP (ResNet-101) · Cityscapes->ACDC (source-free) | mIoU: 43.1 -> 48.4 -> 50.5 | Improved |
| add complementary self-training and propagation to alignment (ablation) | ATP (MiT-B5) · Cityscapes->ACDC (source-free) | mIoU: 56.2 -> 64.4 -> 65.6 | Improved |
| add complementary self-training and propagation to alignment (ablation) | ATP (ResNet-101) · Cityscapes->Cross-City (source-free) | mIoU: 46.5 -> 51.8 -> 53.8 | Improved |
| add information propagation to the black-box source model setting (ablation) | ATP (black-box) · GTA5->Cityscapes (black-box source model) | mIoU: 40.4 -> 44.4 | Improved |
| add augmentation to black-box ATP (ablation) | ATP (black-box) · GTA5->Cityscapes (black-box source model) | mIoU: 44.4 -> 45.3 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- C-SFDA — C-SFDA: A Curriculum Learning Aided Self-Training Framework for Efficient Source Free Domain Adaptation (CVPR 2023) · better on 1 of 1 shared result
- DINE — DINE: Domain Adaptation From Single and Multiple Black-Box Predictors (CVPR 2022) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Constructing Self-Motivated Pyramid Curriculums for Cross-Domain Semantic Segmentation: A Non-Adversarial Approach (ICCV 2019 · similarity 0.92)
- Unsupervised Intra-Domain Adaptation for Semantic Segmentation Through Self-Supervision (CVPR 2020 · similarity 0.92)
- Learning from Scale-Invariant Examples for Domain Adaptation in Semantic Segmentation (ECCV 2020 · similarity 0.91)
- Generalize Then Adapt: Source-Free Domain Adaptive Semantic Segmentation (ICCV 2021 · similarity 0.91)
- Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation (CVPR 2021 · similarity 0.91)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.90)
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.
