Research map / Detection and segmentation / CNN features
Rotation-Sensitive Regression for Oriented Scene Text Detection
Minghui Liao, Zhen Zhu, Baoguang Shi, Gui-song Xia, Xiang Bai
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 inception blocks with multiple receptive-field shapes (ablation) | RRD detector · RCTW-Long | F-measure: 0.6902 -> 0.7532 | Improved |
| Use rotation-sensitive features for regression and classification (ablation) | RRD with inception blocks · RCTW-Long | F-measure: 0.7532 -> 0.8127 | Improved |
| Use rotation-invariant features for classification while keeping rotation-sensitive regression (ablation) | Baseline+inc+rs · RCTW-Long | F-measure: 0.8127 -> 0.8336 | Improved |
| Use RRD rather than the official baseline | scene text detection · RCTW-17 | F-measure: 0.528 -> 0.557 | Improved |
| Use multi-scale testing | RRD · RCTW-17 | F-measure: 0.557 -> 0.670 | Improved |
| Use RRD instead of Baseline+inc (ablation) | line-based text detection · MSRA-TD500 | F-measure: 0.74 -> 0.79 | Improved |
| Use RRD rather than the strongest listed baseline | oriented word detection · ICDAR 2015 | F-measure: 0.782 -> 0.822 | Improved |
| Use RRD+MS rather than Text Detection DL | COCO-Text Challenge, IOU=0.75 · COCO-Text Challenge | F: 0.25 -> 0.36 | Improved |
| Apply RRD to horizontal text detection | RRD · ICDAR 2013 | F-measure: 0.81 -> 0.81 | No clear effect |
| Apply RRD to oriented ship detection | HRSC2016 Level 1 · HRSC2016 | mAP: 75.7 -> 84.3 | Improved |
Closest papers by meaning
- Deep Direct Regression for Multi-Oriented Scene Text Detection (ICCV 2017 · similarity 0.88)
- Embedding Rotation Invariance for Provable Multi-Oriented Scene Text Recognition (ECCV 2026 · similarity 0.87)
- Exploring the Capacity of an Orderless Box Discretization Network for Multi-orientation Scene Text Detection (IJCV 2021 · similarity 0.86)
- Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation (CVPR 2018 · similarity 0.86)
- Adaptive Rotated Convolution for Rotated Object Detection (ICCV 2023 · similarity 0.85)
- Wholly-WOOD: Wholly Leveraging Diversified-Quality Labels for Weakly-Supervised Oriented Object Detection (TPAMI 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.
