Research map / Detection and segmentation / CNN features
SegFix: Model-Agnostic Boundary Refinement for Segmentation
Yuhui Yuan, Jingyi Xie, Xilin Chen, Jingdong Wang
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 SegFix boundary refinement | DeepLabv3 · Cityscapes val | mIoU (%): 79.5 -> 80.5 | Improved |
| add SegFix boundary refinement | DeepLabv3 · Cityscapes val | boundary F-score (%): 56.6 -> 60.9 | Improved |
| add SegFix to HRNet + OCR | HRNet + OCR · Cityscapes test | mIoU (%): 84.2 -> 84.5 | Improved |
| add SegFix to Gated-SCNN | Gated-SCNN · Cityscapes val | boundary F-score (%): 61.4 -> 63.1 | Improved |
| add SegFix to DeepLabv3 on ADE20K | DeepLabv3 · ADE20K val | mIoU (%): 44.8 -> 45.4 | Improved |
| add SegFix to DeepLabv3 on GTA5 | DeepLabv3 · GTA5 test | boundary F-score (%): 50.2 -> 61.7 | Improved |
| add SegFix to PANet | PANet · Cityscapes test (instance) | mask AP (%): 36.4 -> 37.8 | Improved |
| combine GUM guided up-sampling with SegFix | SegFix · Cityscapes val | boundary F-score (%): 60.9 -> 61.6 | Improved |
| replace DenseCRF with SegFix | DenseCRF · Cityscapes val | mIoU (%): 79.7 -> 80.5 | Improved |
| use SegFix refinement instead of two-model ensemble | DeepLabv3+HRNet-W18 ensemble · Cityscapes val | boundary F-score (%): 58.2 -> 60.3 | Improved |
| use larger SegFix backbone HRNet-2x instead of HRNet-W18 (ablation) | SegFix · Cityscapes val | mIoU gain (points): +0.8 -> +1.0 | Improved |
| use boundary width gamma=5 instead of treating all pixels as boundary (ablation) | SegFix · Cityscapes val | boundary F-score gain (points): +3.6 -> +4.2 | Improved |
Later papers that compared with this method
2 later accepted papers report SegFix as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- E-CRF: Embedded Conditional Random Field for Boundary-caused Class Weights Confusion in Semantic Segmentation (ICLR 2023 · better on 1 of 1 shared result)
- Progressive Semantic Segmentation (CVPR 2021 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- PolyTransform — PolyTransform: Deep Polygon Transformer for Instance Segmentation (CVPR 2020) · better on 1 of 1 shared result
- PANet — PANet: Few-Shot Image Semantic Segmentation With Prototype Alignment (ICCV 2019) · 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
- SegRefiner: Towards Model-Agnostic Segmentation Refinement with Discrete Diffusion Process (NeurIPS 2023 · similarity 0.87)
- SharpContour: A Contour-Based Boundary Refinement Approach for Efficient and Accurate Instance Segmentation (CVPR 2022 · similarity 0.86)
- Look Closer To Segment Better: Boundary Patch Refinement for Instance Segmentation (CVPR 2021 · similarity 0.85)
- InverseForm: A Loss Function for Structured Boundary-Aware Segmentation (CVPR 2021 · similarity 0.85)
- Toward Real Ultra Image Segmentation: Leveraging Surrounding Context to Cultivate General Segmentation Model (NeurIPS 2024 · similarity 0.83)
- SegMAN: Omni-scale Context Modeling with State Space Models and Local Attention for Semantic Segmentation (CVPR 2025 · similarity 0.83)
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.
