Research map / Label-efficient and robust learning / Domain adaptation and generalization
Exploring Probabilistic Modeling Beyond Domain Generalization for Semantic Segmentation
I-Hsiang Chen, Hua-En Chang, Wei-Ting Chen, Jenq-Neng Hwang, Sy-Yen Kuo
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 PDAF to DeepLabV3Plus ResNet-50 | DeepLabV3Plus · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 40.62 -> 49.41 | Improved |
| add PDAF to Mask2Former Swin-L | Mask2Former · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 58.13 -> 61.08 | Improved |
| add PDAF to Mask2Former Swin-T on ACDC | Mask2Former · ACDC (Cityscapes-trained) | mIoU (avg over unseen domains): 53.15 -> 62.97 | Improved |
| add the Latent Prior Extractor (variational posterior) (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 45.19 -> 49.41 | Improved |
| add the Domain Compensation Module (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 47.82 -> 49.41 | Improved |
| add the Diffusion Prior Estimator (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 47.64 -> 49.41 | Improved |
| add semantic consistency loss on top of task loss (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 45.83 -> 46.80 | Improved |
| add prior constraint loss on top of task loss (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 45.83 -> 48.82 | Improved |
| add semantic consistency loss to task + prior loss (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 48.82 -> 49.41 | Improved |
| increase LDP channels from 4 to 1024 (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 49.41 -> 45.69 | Hurt |
| reduce LDP channels from 4 to 1 (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 49.41 -> 48.74 | Hurt |
| increase DPE diffusion steps from 4 to 32 (ablation) | PDAF · Cityscapes-trained, unseen B/M/G/S | mIoU (avg over unseen domains): 49.41 -> 49.48 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- WildNet — WildNet: Learning Domain Generalized Semantic Segmentation From the Wild (CVPR 2022) · better on 1 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
- Collaborating Foundation Models for Domain Generalized Semantic Segmentation (CVPR 2024 · similarity 0.90)
- Better to Teach than to Give: Domain Generalized Semantic Segmentation via Agent Queries with Diffusion Model Guidance (ICML 2025 · similarity 0.89)
- Exploring Semantic Consistency and Style Diversity for Domain Generalized Semantic Segmentation (AAAI 2025 · similarity 0.89)
- FisherTune: Fisher-Guided Robust Tuning of Vision Foundation Models for Domain Generalized Segmentation (CVPR 2025 · similarity 0.89)
- A Simple Recipe for Language-guided Domain Generalized Segmentation (CVPR 2024 · similarity 0.89)
- Stronger Fewer & Superior: Harnessing Vision Foundation Models for Domain Generalized Semantic Segmentation (CVPR 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.
