Research map / Detection and segmentation / CNN features
CBNet: A Plug-and-Play Network for Segmentation-Based Scene Text Detection
Xi Qing Zhao, Wei Feng, Zheng Zhang, Jingjing Lv, Xin Zhu, Zhangang Lin, Jinghe Hu, Jingping Shao
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 global and local context-aware module (GL-CAM) (ablation) | PAN-1s · CTW1500 | F-measure (%): 83.7 -> 85.3 | Improved |
| add global and local context-aware module (GL-CAM) (ablation) | PAN-4s · CTW1500 | F-measure (%): 83.5 -> 84.7 | Improved |
| add global context module alone (ablation) | PAN-4s · CTW1500 | F-measure (%): 83.5 -> 84.1 | Improved |
| fuse local with global context module (ablation) | PAN-1s · CTW1500 | F-measure (%): 84.4 -> 85.3 | Improved |
| use GL-CAM instead of non-local, GCN or PCAA modules (ablation) | PAN-4s · CTW1500 (from scratch) | F-measure (%): 82.1 -> 83.3 | Improved |
| use boundary-guided kernel expansion instead of DBNet post-processing (ablation) | PAN-1s+GL-CAM · CTW1500 | F-measure (%): 82.4 -> 86.0 | Improved |
| use boundary-guided kernel expansion instead of pixel aggregation (ablation) | PAN-1s+GL-CAM · CTW1500 | F-measure (%): 85.9 -> 86.0 | Improved |
| use boundary-guided expansion instead of pixel aggregation (post-processing time) (ablation) | PAN-1s+GL-CAM · CTW1500 | post-processing time (ms): 110.7 -> 9.2 | Improved |
| add CBN plug-in | PSENet · CTW1500 (from scratch) | F-measure (%): 78.9 -> 80.3 | Improved |
| add CBN plug-in | PAN · CTW1500 (from scratch) | F-measure (%): 82.0 -> 83.3 | Improved |
| add CBN plug-in | DBNet · CTW1500 (from scratch) | F-measure (%): 80.1 -> 81.0 | Improved |
| add CBN plug-in | PAN · Total-Text (from scratch) | F-measure (%): 83.5 -> 84.5 | Improved |
Closest papers by meaning
- A Text Detection System for Natural Scenes With Convolutional Feature Learning and Cascaded Classification (CVPR 2016 · similarity 0.89)
- Kernel Adaptive Convolution for Scene Text Detection via Distance Map Prediction (CVPR 2024 · similarity 0.89)
- Scene Text Detection with Supervised Pyramid Context Network (AAAI 2019 · similarity 0.88)
- ContourNet: Taking a Further Step Toward Accurate Arbitrary-Shaped Scene Text Detection (CVPR 2020 · similarity 0.88)
- PixelLink: Detecting Scene Text via Instance Segmentation (AAAI 2018 · similarity 0.88)
- Learning Shape-Aware Embedding for Scene Text Detection (CVPR 2019 · similarity 0.88)
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.
