Research map / Detection and segmentation / CNN features
Mask TextSpotter v3: Segmentation Proposal Network for Robust Scene Text Spotting
Minghui Liao, Guan Pang, Jing Huang, Tal Hassner, 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 hard RoI masking (Direct-hard) (ablation) | Mask TextSpotter v3 · Total-Text (no lexicon) | F-measure: 67.3 -> 71.2 | Improved |
| add hard RoI masking (Direct-hard) (ablation) | Mask TextSpotter v3 · Total-Text (full lexicon) | F-measure: 76.2 -> 78.4 | Improved |
| add hard RoI masking (Direct-hard) (ablation) | Mask TextSpotter v3 · IC15 (strong lexicon) | F-measure: 81.0 -> 83.3 | Improved |
| use binary hard mask instead of soft probability mask (direct) (ablation) | Mask TextSpotter v3 · Total-Text (no lexicon) | F-measure: 69.1 -> 71.2 | Improved |
| apply the mask directly instead of through extra layers (hard) (ablation) | Mask TextSpotter v3 · Total-Text (no lexicon) | F-measure: 68.4 -> 71.2 | Improved |
| add indirect soft masking via learned mask layers (ablation) | Mask TextSpotter v3 · Total-Text (no lexicon) | F-measure: 67.3 -> 65.8 | Hurt |
| add soft RoI masking (Direct-soft) in place of none (ablation) | Mask TextSpotter v3 · Total-Text (full lexicon) | F-measure: 76.2 -> 76.0 | No clear effect |
| replace RPN with Segmentation Proposal Network (v2 to v3) | Mask TextSpotter · RoIC13 detection (rotation 45 deg) | F-measure: 62.2 -> 84.2 | Improved |
| replace RPN with Segmentation Proposal Network (v2 to v3) | Mask TextSpotter · RoIC13 end-to-end (rotation 45 deg) | F-measure: 54.2 -> 76.1 | Improved |
| replace RPN with Segmentation Proposal Network (v2 to v3) | Mask TextSpotter · MSRA-TD500 | F-measure: 74.2 -> 83.5 | Improved |
| replace RPN with Segmentation Proposal Network (v2 to v3) | Mask TextSpotter · Total-Text (no lexicon) | F-measure: 65.3 -> 71.2 | Improved |
| replace RPN with Segmentation Proposal Network (v2 to v3) | Mask TextSpotter · IC15 end-to-end (generic lexicon) | F-measure: 73.5 -> 74.2 | Improved |
Later papers that compared with this method
3 later accepted papers report MaskTextSpotter v3 as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- ESTextSpotter: Towards Better Scene Text Spotting with Explicit Synergy in Transformer (ICCV 2023 · better on 1 of 1 shared result)
- SwinTextSpotter: Scene Text Spotting via Better Synergy Between Text Detection and Text Recognition (CVPR 2022 · better on 3 of 4 shared results)
- CentripetalText: An Efficient Text Instance Representation for Scene Text Detection (NeurIPS 2021 · better on 2 of 2 shared results)
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.
- TextDragon — TextDragon: An End-to-End Framework for Arbitrary Shaped Text Spotting (ICCV 2019) · better on 2 of 2 shared results
- ABCNet — ABCNet: Real-Time Scene Text Spotting With Adaptive Bezier-Curve Network (CVPR 2020) · better on 1 of 1 shared result
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
- Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes (ECCV 2018 · similarity 0.91)
- SPTS v2: Single-Point Scene Text Spotting (TPAMI 2023 · similarity 0.90)
- CPN: Complementary Proposal Network for Unconstrained Text Detection (AAAI 2024 · similarity 0.89)
- Towards Unconstrained End-to-End Text Spotting (ICCV 2019 · similarity 0.89)
- ContourNet: Taking a Further Step Toward Accurate Arbitrary-Shaped Scene Text Detection (CVPR 2020 · similarity 0.89)
- SemiETS: Integrating Spatial and Content Consistencies for Semi-Supervised End-to-end Text Spotting (CVPR 2025 · 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.
