Research map / Detection and segmentation / Open-set and few-shot detection
TransFGU: A Top-down Approach to Fine-Grained Unsupervised Semantic Segmentation
Zhaoyuan Yin, Pichao Wang, Fan Wang, Xianzhe Xu, Hanling Zhang, Hao Li, Rong Jin
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 teacher-student bootstrapping iterations after training the student once (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 8.66 -> 11.93 | Improved |
| train a student on the initial pseudo labels (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 5.05 -> 8.66 | Improved |
| use ViT with Segmenter decoder instead of R50+FPN encoder (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 9.21 -> 11.93 | Improved |
| use Segmenter decoder instead of conv classifier on ViT (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 7.96 -> 11.93 | Improved |
| add peer loss to CE (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 10.49 -> 10.54 | Improved |
| add uncertainty loss to peer loss (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 10.54 -> 11.24 | Improved |
| add diversity loss to peer loss (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 10.54 -> 10.96 | Improved |
| combine peer, uncertainty and diversity losses (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 11.24 -> 11.93 | Improved |
| use DINO ViT-S/8 instead of ViT-S/16 as prior (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 9.43 -> 11.93 | Improved |
| use DINO instead of MoCoV3 as the prior (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 2.32 -> 9.43 | Improved |
| separate foreground and background patches before clustering (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 9.52 -> 11.93 | Improved |
| use ground-truth crops instead of sliding windows (ablation) | TransFGU · COCO-Stuff-171 | mIoU: 11.93 -> 12.16 | 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.
- PiCIE — PiCIE: Unsupervised Semantic Segmentation Using Invariance and Equivariance in Clustering (CVPR 2021) · better on 4 of 4 shared results
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
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.90)
- Integrating Low-Level Visual Cues for Enhanced Unsupervised Semantic Segmentation (AAAI 2025 · similarity 0.90)
- Segment Anything without Supervision (NeurIPS 2024 · similarity 0.89)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.89)
- Unsupervised Universal Image Segmentation (CVPR 2024 · similarity 0.89)
- Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling (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.
