Research map / Detection and segmentation / CNN features
Deep Hough Transform for Semantic Line Detection
Qi Han, Kai Zhao, Jun Xu, Ming-Ming Cheng
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 multi-scale DHT (ablation) | DHT baseline (ResNet50) · SEL | F-measure: 0.664 -> 0.758 | Improved |
| add context-aware line detector (CTX) (ablation) | DHT baseline (ResNet50) · SEL | F-measure: 0.664 -> 0.771 | Improved |
| add CTX on top of DHT+MS (ablation) | DHT + MS · SEL | F-measure: 0.758 -> 0.786 | Improved |
| add MS on top of DHT+CTX (ablation) | DHT + CTX · SEL | F-measure: 0.771 -> 0.786 | Improved |
| add edge-guided refinement with delta_r=5 (ablation) | DHT ResNet50 · SEL | F-measure: 0.7861 -> 0.7880 | Improved |
| increase delta_r from 5 to 9 in edge-guided refinement (ablation) | DHT+ER · SEL | F-measure: 0.7880 -> 0.7877 | No clear effect |
| add edge-guided refinement (ablation) | DHT VGG16 · SEL | F@0.95: 0.380 -> 0.439 | Improved |
| add edge-guided refinement (ablation) | DHT ResNet50 · SEL | F@0.95: 0.420 -> 0.461 | Improved |
| add edge-guided refinement (ablation) | DHT ResNet50 · NKL | F-measure: 0.719 -> 0.725 | Improved |
| add edge-guided refinement (ablation) | DHT VGG16 · NKL | F-measure: 0.706 -> 0.711 | Improved |
| replace VGG16 backbone with ResNet50 | DHT · SEL | F-measure (EA): 0.765 -> 0.786 | Improved |
| replace SLNet-iter10 with deep Hough transform | SLNet · SEL | F-measure (EA): 0.745 -> 0.786 | Improved |
Closest papers by meaning
- Deep Hough-Transform Line Priors (ECCV 2020 · similarity 0.92)
- DeepLSD: Line Segment Detection and Refinement With Deep Image Gradients (CVPR 2023 · similarity 0.89)
- A Comprehensive Review of Image Line Segment Detection and Description: Taxonomies, Comparisons, and Challenges (TPAMI 2024 · similarity 0.89)
- SOLD2: Self-Supervised Occlusion-Aware Line Description and Detection (CVPR 2021 · similarity 0.89)
- ScaleLSD: Scalable Deep Line Segment Detection Streamlined (CVPR 2025 · similarity 0.87)
- Semantic Line Combination Detector (CVPR 2024 · 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.
