Research map / Image generation / Fast diffusion sampling
SegRefiner: Towards Model-Agnostic Segmentation Refinement with Discrete Diffusion Process
Mengyu Wang, Henghui Ding, Jun Hao Liew, Jiajun Liu, Yao Zhao, Yunchao Wei
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 discrete diffusion sampling to iterative refinement (vs none) (ablation) | SegRefiner · BIG | mBA: 75.03 -> 77.64 | Improved |
| add discrete diffusion sampling to iterative refinement (vs none) (ablation) | SegRefiner · BIG | IoU: 94.10 -> 94.85 | Improved |
| iterate with previous mask as input but no diffusion sampling (ablation) | SegRefiner · BIG | mBA: 75.03 -> 75.10 | No clear effect |
| add local refinement to global refinement (ablation) | SegRefiner · BIG | mBA: 60.58 -> 77.64 | Improved |
| add global refinement to local refinement (ablation) | SegRefiner · BIG | IoU: 89.73 -> 94.85 | Improved |
| add local refinement to global refinement (ablation) | SegRefiner · BIG | IoU: 92.43 -> 94.85 | Improved |
| input size 128x128 to 256x256 (ablation) | SegRefiner · BIG | IoU: 93.10 -> 94.85 | Improved |
| input size 256x256 to 512x512 (ablation) | SegRefiner · BIG | mBA: 77.64 -> 77.56 | No clear effect |
| increase iteration steps from 1 to 6 (ablation) | SegRefiner · COCO val (LVIS), Mask R-CNN R50 | Boundary AP: 29.9 -> 32.6 | Improved |
| increase iteration steps from 5 to 6 (ablation) | SegRefiner · COCO val (LVIS), Mask R-CNN R50 | Mask AP: 41.9 -> 41.9 | No clear effect |
| refine Mask R-CNN R50 with SegRefiner instead of BPR | Mask R-CNN R50 · COCO val (LVIS) | Boundary AP: 30.4 -> 32.6 | Improved |
| apply SegRefiner to SOLO | SOLO · COCO val (LVIS) | Boundary AP: 24.7 -> 31.3 | Improved |
Closest papers by meaning
- Seg-VAR:Image Segmentation with Visual Autoregressive Modeling (NeurIPS 2025 · similarity 0.88)
- $\mathbf{D^{3}S^{2}}$: Diffusion-Guided Dataset Distillation for Semantic Segmentation (NeurIPS 2026 · similarity 0.86)
- High-Precision Dichotomous Image Segmentation via Probing Diffusion Capacity (ICLR 2025 · similarity 0.86)
- MaskFactory: Towards High-quality Synthetic Data Generation for Dichotomous Image Segmentation (NeurIPS 2024 · similarity 0.86)
- SemFlow: Binding Semantic Segmentation and Image Synthesis via Rectified Flow (NeurIPS 2024 · similarity 0.86)
- SMITE: Segment Me In TimE (ICLR 2025 · similarity 0.86)
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.
