Research map / Detection and segmentation / CNN features
TP-LSD: Tri-Points Based Line Segment Detector
Siyu Huang, Fangbo Qin, Pengfei Xiong, Ning Ding, Yijia He, Xiao Liu
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 line-map segmentation branch (LSB) (ablation) | TP extraction branch · Wireframe | F^H: 0.782 -> 0.808 | Improved |
| add line-map segmentation branch (LSB) (ablation) | TP extraction branch · Wireframe | LAP: 58.6 -> 61.4 | Improved |
| add feature aggregate module (FAM) (ablation) | TP-LSD with LSB · Wireframe | LAP: 61.4 -> 59.2 | Hurt |
| add feature aggregate module (FAM) (ablation) | TP-LSD with LSB · Wireframe | sAP^10: 59.2 -> 60.4 | Improved |
| add feature aggregate module (FAM) (ablation) | TP-LSD with LSB · Wireframe | F^H: 0.808 -> 0.810 | Improved |
| add mixture convolution module (MCM) (ablation) | TP-LSD with LSB+FAM · Wireframe | LAP: 59.2 -> 60.9 | Improved |
| add mixture convolution module (MCM) (ablation) | TP-LSD with LSB+FAM · Wireframe | sAP^10: 60.4 -> 61.3 | Improved |
| add PFM with alpha=0.5 (ablation) | TP-LSD with LSB+FAM+MCM · Wireframe | F^H: 0.810 -> 0.816 | Improved |
| add PFM with alpha=0.5 (ablation) | TP-LSD with LSB+FAM+MCM · Wireframe | LAP: 60.9 -> 60.6 | Hurt |
| use PFM alpha=0.5 instead of alpha=1 (ablation) | TP-LSD with LSB+FAM+MCM · Wireframe | LAP: 60.1 -> 60.6 | Improved |
| add rotation and scale augmentation to flips (ablation) | TP-LSD-Res34 · Wireframe | LAP: 60.1 -> 60.6 | Improved |
| use Hourglass backbone instead of ResNet34 at 512 input | TP-LSD · Wireframe | LAP: 61.3 -> 55.1 | Hurt |
Closest papers by meaning
- Towards Light-Weight and Real-Time Line Segment Detection (AAAI 2022 · similarity 0.90)
- ScaleLSD: Scalable Deep Line Segment Detection Streamlined (CVPR 2025 · similarity 0.90)
- Learning Attraction Field Representation for Robust Line Segment Detection (CVPR 2019 · similarity 0.88)
- DeepLSD: Line Segment Detection and Refinement With Deep Image Gradients (CVPR 2023 · similarity 0.88)
- SOLD2: Self-Supervised Occlusion-Aware Line Description and Detection (CVPR 2021 · similarity 0.86)
- Improving Transformer Based Line Segment Detection with Matched Predicting and Re-ranking (AAAI 2025 · similarity 0.85)
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.
