Research map / Vision-language and multimodal / Human and embodied modeling
Scene Text Recognition from Two-Dimensional Perspective
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 |
|---|---|---|---|
| Represent and classify characters in two-dimensional pixel space instead of compressing text features into a one-dimensional sequence | CA-FCN · CUTE | accuracy: 3.1 percents improvement | Helped |
| Fine-tune with extra synthetic data | CA-FCN · SVT (50-word lexicon) | accuracy: 98.5 -> 98.8 | Helped |
| Add character attention (ablation) | CA-FCN baseline · IIIT | accuracy: 90.5 -> 91.0 | Helped |
| Add character attention (ablation) | CA-FCN baseline · IIIT-p | accuracy: 87.0 -> 86.7 | Hurt |
| Add deformable convolution (ablation) | CA-FCN baseline · IIIT-p | accuracy: 87.0 -> 87.6 | Helped |
| Add character attention and deformable convolution (ablation) | CA-FCN baseline · IIIT-p | accuracy: 87.0 -> 89.3 | Helped |
| Add deformable convolution (ablation) | CA-FCN baseline · IC13-ex | accuracy: 83.2 -> 87.4 | Helped |
| Add character attention and deformable convolution (ablation) | CA-FCN baseline · IC13-ex | accuracy: 83.2 -> 87.2 | Helped |
| Add deformable convolution (ablation) | CA-FCN baseline · IC13-r-ex | accuracy: 82.3 -> 84.2 | Helped |
| Use a two-dimensional character prediction approach rather than sequence-based recognition | CA-FCN · IIIT-p | accuracy gap ratio: 6.4% -> 2.6% | Helped |
Closest papers by meaning
- A Text Attention Network for Spatial Deformation Robust Scene Text Image Super-Resolution (CVPR 2022 · similarity 0.89)
- Char-Net: A Character-Aware Neural Network for Distorted Scene Text Recognition (AAAI 2018 · similarity 0.88)
- Character Region Awareness for Text Detection (CVPR 2019 · similarity 0.88)
- TextScanner: Reading Characters in Order for Robust Scene Text Recognition (AAAI 2020 · similarity 0.88)
- OTE: Exploring Accurate Scene Text Recognition Using One Token (CVPR 2024 · similarity 0.88)
- Kernel Adaptive Convolution for Scene Text Detection via Distance Map Prediction (CVPR 2024 · similarity 0.87)
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.