Research map / Vision-language and multimodal / Human and embodied modeling
Robust Scene Text Recognition With Automatic Rectification
Check 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 |
|---|---|---|---|
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IIIT5K (None) | word accuracy: 79.7 -> 81.9 | Helped |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IIIT5K (50) | word accuracy: 96.5 -> 96.2 | Hurt |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IIIT5K (1k) | word accuracy: 92.8 -> 93.8 | Helped |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · SVT (50) | word accuracy: 96.1 -> 95.5 | Hurt |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · SVT (None) | word accuracy: 81.5 -> 81.9 | Helped |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IC03 (50) | word accuracy: 97.8 -> 98.3 | Helped |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IC03 (Full) | word accuracy: 96.4 -> 96.2 | Hurt |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IC03 (50k) | word accuracy: 93.7 -> 94.8 | Helped |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IC03 (None) | word accuracy: 88.7 -> 90.1 | Helped |
| Add TPS-based STN rectification to the SRN (ablation) | RARE · IC13 (None) | word accuracy: 87.5 -> 88.6 | Helped |
| Use RARE rather than Shi et al. | scene text recognition · SVT-Perspective (Full) | word accuracy: 72.6 -> 77.4 | Helped |
| Use RARE rather than Shi et al. | scene text recognition · SVT-Perspective (None) | word accuracy: 66.8 -> 71.8 | Helped |
Closest papers by meaning
- Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition (AAAI 2019 · similarity 0.90)
- AON: Towards Arbitrarily-Oriented Text Recognition (CVPR 2018 · similarity 0.88)
- ESIR: End-To-End Scene Text Recognition via Iterative Image Rectification (CVPR 2019 · similarity 0.88)
- OTE: Exploring Accurate Scene Text Recognition Using One Token (CVPR 2024 · similarity 0.85)
- Out of Length Text Recognition with Sub-String Matching (AAAI 2025 · similarity 0.85)
- Boosting Semi-Supervised Scene Text Recognition via Viewing and Summarizing (NeurIPS 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.