Research map / Detection and segmentation / CNN features
Learning Regional Attraction for Line Segment Detection
Nan Xue, Song Bai, Fudong Wang, Gui-Song Xia, Tianfu Wu, Liangpei Zhang, Philip H. S. Torr
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 |
|---|---|---|---|
| train and test at 512x512 instead of 320x320 (ablation) | AFM a-trous, SGD, no outlier removal · Wireframe | F-measure: 0.773 -> 0.794 | Improved |
| train and test at 512x512 instead of 320x320 (ablation) | AFM a-trous, SGD, no outlier removal · YorkUrban | F-measure: 0.647 -> 0.660 | Improved |
| add outlier removal (ablation) | AFM a-trous, SGD, 320x320 · Wireframe | F-measure: 0.774 -> 0.807 | Improved |
| add outlier removal (ablation) | AFM a-trous, SGD, 320x320 · YorkUrban | F-measure: 0.647 -> 0.663 | Improved |
| add outlier removal (ablation) | AFM a-trous, Adam, 512x512 · Wireframe | F-measure: 0.802 -> 0.831 | Improved |
| use Adam instead of SGD (ablation) | AFM a-trous, 320x320, no outlier removal · Wireframe | F-measure: 0.774 -> 0.792 | Improved |
| use Adam instead of SGD (ablation) | AFM a-trous, 320x320, outlier removal · Wireframe | F-measure: 0.807 -> 0.823 | Improved |
| use Adam instead of SGD (ablation) | AFM a-trous, 512x512, outlier removal · YorkUrban | F-measure: 0.674 -> 0.680 | Improved |
| replace U-Net with a-trous Residual U-Net (ablation) | AFM · Wireframe | F-measure: 0.753 -> 0.774 | Improved |
| add outlier removal, Adam to AFM (AFM++) (ablation) | AFM a-trous · YorkUrban | F-measure: 0.647 -> 0.672 | Improved |
| increase resolution to 512x512 (inference speed) (ablation) | AFM a-trous, SGD, no outlier removal · Wireframe | FPS: 6.6 -> 2.3 | Hurt |
| add outlier removal (inference speed) (ablation) | AFM a-trous, 320x320 · Wireframe | FPS: 6.6 -> 8.0 | Improved |
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.
- MCMLSD — MCMLSD: A Dynamic Programming Approach to Line Segment Detection (CVPR 2017) · better on 3 of 3 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
- Learning Attraction Field Representation for Robust Line Segment Detection (CVPR 2019 · similarity 0.93)
- MCMLSD: A Dynamic Programming Approach to Line Segment Detection (CVPR 2017 · similarity 0.87)
- DeepLSD: Line Segment Detection and Refinement With Deep Image Gradients (CVPR 2023 · similarity 0.86)
- ScaleLSD: Scalable Deep Line Segment Detection Streamlined (CVPR 2025 · similarity 0.86)
- Improving Transformer Based Line Segment Detection with Matched Predicting and Re-ranking (AAAI 2025 · similarity 0.84)
- MFP: Making Full Use of Probability Maps for Interactive Image Segmentation (CVPR 2024 · 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.
