Research map / Label-efficient and robust learning / Domain adaptation and generalization
Multi-source Domain Adaptation for 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 MADAN's multi-source alignment and aggregation framework rather than source-combined CyCADA | Multi-source semantic segmentation · GTA+SYNTHIA → Cityscapes | mIoU: 37.3 -> 41.4 | Helped |
| Use MADAN's multi-source alignment and aggregation framework rather than source-combined CyCADA | Multi-source semantic segmentation · GTA+SYNTHIA → BDDS | mIoU: 33.7 -> 36.3 | Helped |
| Add sub-domain aggregation discriminator (SAD) (ablation) | MADAN · GTA+SYNTHIA → Cityscapes | mIoU: 30.0 -> 36.4 | Helped |
| Add cross-domain cycle discriminator (CCD) (ablation) | MADAN · GTA+SYNTHIA → Cityscapes | mIoU: 30.0 -> 33.1 | Helped |
| Add sub-domain aggregation discriminator (SAD) (ablation) | MADAN · GTA+SYNTHIA → BDDS | mIoU: 24.6 -> 31.2 | Helped |
| Add cross-domain cycle discriminator (CCD) (ablation) | MADAN · GTA+SYNTHIA → BDDS | mIoU: 24.6 -> 29.3 | Helped |
| Use dynamic semantic consistency (DSC) instead of original semantic consistency (SC) (ablation) | CycleGAN and CyCADA · GTA and SYNTHIA → Cityscapes and BDDS | mIoU: 32.7 -> 38.1; 38.7 -> 40.0; 27.3 -> 30.5; 29.2 -> 31.8; 31.1 -> 32.9; 32.0 -> 35.5; 21.8 -> 22.5; 23.4 -> 24.0 | Helped |
| Add feature-level alignment to SAD+CCD+DSC (ablation) | MADAN · GTA+SYNTHIA → Cityscapes | mIoU: 40.2 -> 41.4 | Helped |
| Add feature-level alignment to SAD+CCD+DSC (ablation) | MADAN · GTA+SYNTHIA → BDDS | mIoU: 35.3 -> 36.3 | Helped |
| Combine SAD and CCD rather than use SAD alone (ablation) | MADAN · GTA+SYNTHIA → Cityscapes | mIoU: 36.4 -> 37.5 | Helped |
| Combine SAD and CCD rather than use SAD alone (ablation) | MADAN · GTA+SYNTHIA → BDDS | mIoU: 31.2 -> 31.8 | Helped |
Closest papers by meaning
- Multi-Source Domain Adaptation With Collaborative Learning for Semantic Segmentation (CVPR 2021 · similarity 0.92)
- Pixel-Level Cycle Association: A New Perspective for Domain Adaptive Semantic Segmentation (NeurIPS 2020 · similarity 0.91)
- Learning to Adapt Structured Output Space for Semantic Segmentation (CVPR 2018 · similarity 0.91)
- An Adversarial Perturbation Oriented Domain Adaptation Approach for Semantic Segmentation (AAAI 2020 · similarity 0.91)
- Learning From Synthetic Data: Addressing Domain Shift for Semantic Segmentation (CVPR 2018 · similarity 0.90)
- Learning to Adapt via Latent Domains for Adaptive Semantic Segmentation (NeurIPS 2021 · 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.