Research map / Detection and segmentation / CNN features
Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes
Pengyuan Lyu, Minghui Liao, Cong Yao, Wenhao Wu, 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 the character-map subnet (joint detection and recognition training) (ablation) | Mask TextSpotter · ICDAR2013 detection | F-measure: 91.0 -> 91.7 | Improved |
| add the character-map subnet (joint detection and recognition training) (ablation) | Mask TextSpotter · ICDAR2015 detection | F-measure: 83.4 -> 86.0 | Improved |
| add a few real-world character annotations (ablation) | Mask TextSpotter · ICDAR2015 end-to-end, strong | F-measure: 76.6 -> 79.3 | Improved |
| add a few real-world character annotations (ablation) | Mask TextSpotter · ICDAR2013 end-to-end | F-measure: 0.7%-2.3% | Improved |
| use weighted edit distance instead of original edit distance (ablation) | Mask TextSpotter · ICDAR2015 end-to-end, generic | F-measure: 56.7 -> 62.4 | Improved |
| use weighted edit distance instead of original edit distance (ablation) | Mask TextSpotter · ICDAR2015 end-to-end, weak | F-measure: 67.1 -> 73.0 | Improved |
| increase input scale from 720 to 1600 | Mask TextSpotter · ICDAR2015 end-to-end, strong | F-measure: 71.3 -> 79.3 | Improved |
| increase input scale from 720 to 1600 | Mask TextSpotter · ICDAR2015 | FPS: 6.9 -> 2.6 | Hurt |
| use Mask TextSpotter instead of Liao et al. on curved text | Liao et al. · Total-Text end-to-end, no lexicon | F-measure: 36.3 -> 52.9 | Improved |
| use Mask TextSpotter instead of Li et al. on ICDAR2013 generic lexicon | Li et al. · ICDAR2013 word spotting, strong | F-measure: 94.2 -> 92.5 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Textboxes — TextBoxes: A Fast Text Detector with a Single Deep Neural Network (AAAI 2017) · better on 3 of 3 shared results
- Deep text spotter — Deep TextSpotter: An End-To-End Trainable Scene Text Localization and Recognition Framework (ICCV 2017) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Towards End-To-End Text Spotting With Convolutional Recurrent Neural Networks (ICCV 2017 · similarity 0.92)
- Mask TextSpotter v3: Segmentation Proposal Network for Robust Scene Text Spotting (ECCV 2020 · similarity 0.91)
- Towards Unconstrained End-to-End Text Spotting (ICCV 2019 · similarity 0.91)
- Towards End-to-End Text Spotting in Natural Scenes (TPAMI 2021 · similarity 0.90)
- Text Perceptron: Towards End-to-End Arbitrary-Shaped Text Spotting (AAAI 2020 · similarity 0.90)
- Bridging the Gap Between End-to-End and Two-Step Text Spotting (CVPR 2024 · 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.
