Research map / Detection and segmentation / CNN features
TextScanner: Reading Characters in Order for Robust Scene Text Recognition
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use separate geometry and class branches with character localization (ablation) | TextScanner · IC15 | accuracy: 72.6 -> 75.8 | Helped |
| Use separate geometry and class branches with character localization (ablation) | TextScanner · CUTE80 | accuracy: 81.9 -> 82.7 | Helped |
| Use order maps for word formation and sequential character decoding (ablation) | TextScanner with geometry branch · IC15 | accuracy: 75.8 -> 80.1 | Helped |
| Use order maps for word formation and sequential character decoding (ablation) | TextScanner with geometry branch · SVTP | accuracy: 75.6 -> 83.7 | Helped |
| Use order maps for word formation and sequential character decoding (ablation) | TextScanner with geometry branch · CUTE80 | accuracy: 82.7 -> 85.2 | Helped |
| Use mutual supervision between class and geometry predictions | TextScanner trained on SynthText · IC15 | accuracy: 79.6 -> 80.1 | Helped |
| Use mutual supervision between class and geometry predictions | TextScanner trained on SynthText · CUTE80 | accuracy: 83.5 -> 85.2 | Helped |
| Fine-tune with real training images using sequence-level annotations | TextScanner · IC15 | accuracy: 79.4 -> 83.5 | Helped |
| Fine-tune with real training images using sequence-level annotations | TextScanner · SVTP | accuracy: 86.4 -> 84.8 | Mixed |
| Separate character-order prediction from character classification | TextScanner · MLT-17 Chinese | accuracy: 59.2 -> 64.1 | Helped |
Closest papers by meaning
- Scene Text Recognition from Two-Dimensional Perspective (AAAI 2019 · similarity 0.88)
- Arbitrary Reading Order Scene Text Spotter with Local Semantics Guidance (AAAI 2025 · similarity 0.88)
- OTE: Exploring Accurate Scene Text Recognition Using One Token (CVPR 2024 · similarity 0.87)
- LayoutFormer: Hierarchical Text Detection Towards Scene Text Understanding (CVPR 2024 · similarity 0.86)
- Boosting Semi-Supervised Scene Text Recognition via Viewing and Summarizing (NeurIPS 2024 · similarity 0.86)
- Text Grouping Adapter: Adapting Pre-trained Text Detector for Layout Analysis (CVPR 2024 · 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.