Research map / Detection and segmentation / Weakly supervised detection and segmentation
SmooSeg: Smoothness Prior for Unsupervised Semantic Segmentation
Mengcheng Lan, Xinjiang Wang, Yiping Ke, Jiaxing Xu, Litong Feng, Wayne 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 |
|---|---|---|---|
| remove data term E_data (ablation) | SmooSeg · COCOStuff | pixel accuracy: 10.2% | Improved |
| remove data term E_data (ablation) | SmooSeg · COCOStuff | mIoU: 1.2% | Improved |
| remove across-image smoothness term E_smooth^across (ablation) | SmooSeg · COCOStuff | pixel accuracy: 27.1% | Improved |
| remove across-image smoothness term E_smooth^across (ablation) | SmooSeg · COCOStuff | mIoU: 16.3% | Improved |
| remove entire smoothness term E_smooth (ablation) | SmooSeg · COCOStuff | accuracy and mIoU | Improved |
| add CRF post-processing (ablation) | SmooSeg · COCOStuff | pixel accuracy: 60.6 -> 63.2 | Improved |
| add CRF post-processing (ablation) | SmooSeg · Cityscapes | pixel accuracy: 79.8 -> 82.8 | Improved |
| add CRF post-processing (ablation) | SmooSeg · Potsdam-3 | mIoU: 68.4 -> 70.3 | Improved |
| add CRF post-processing (ablation) | STEGO · Cityscapes | pixel accuracy: 63.5 -> 69.8 | Improved |
| add CRF post-processing (ablation) | SmooSeg · Average over three datasets | pixel accuracy: 73.9 -> 76.2 | Improved |
| add CRF post-processing (ablation) | STEGO · Average over three datasets | pixel accuracy: 61.4 -> 65.0 | Improved |
| lower temperature tau (<=0.1 versus >=0.2) (ablation) | SmooSeg · COCOStuff | accuracy and mIoU | Mixed |
Closest papers by meaning
- Unsupervised Universal Image Segmentation (CVPR 2024 · similarity 0.88)
- Improving Semi-Supervised Semantic Segmentation with Sliced-Wasserstein Feature Alignment and Uniformity (CVPR 2025 · similarity 0.88)
- Integrating Low-Level Visual Cues for Enhanced Unsupervised Semantic Segmentation (AAAI 2025 · similarity 0.87)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.87)
- Uncertainty-aware Fine-tuning of Segmentation Foundation Models (NeurIPS 2024 · similarity 0.87)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · 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.
