Research map / Detection and segmentation / CNN features
SwinTextSpotter v2: Towards Better Synergy for Scene Text Spotting
Mingxin Huang, Dezhi Peng, Hongliang Li, Zhenghao Peng, Chongyu Liu, Dahua Lin, Yuliang Liu, Xiang Bai, Lianwen Jin
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 Box Selection Schedule (ablation) | Baseline+TRS · Total-Text | E2E H-mean: 63.3 -> 67.9 | Improved |
| add Box Selection Schedule (ablation) | Baseline+TRS · ICDAR 2015 Generic | H-mean: 61.4 -> 65.4 | Improved |
| add Recognition Alignment (ablation) | Baseline+TRS · Total-Text | E2E H-mean: 63.3 -> 65.1 | Improved |
| add Recognition Alignment (ablation) | Baseline+TRS · ICDAR 2015 Generic | H-mean: 61.4 -> 64.7 | Improved |
| add RA on top of TRS+BSS (ablation) | SwinTextSpotter v2 · Total-Text | E2E H-mean: 67.9 -> 71.3 | Improved |
| add Recognition Conversion (ablation) | SwinTextSpotter v2 · Total-Text | E2E H-mean: 68.1 -> 71.3 | Improved |
| add TLSAM (ablation) | SwinTextSpotter v2 · Total-Text | E2E H-mean: 70.6 -> 71.3 | Improved |
| add TLSAM (ablation) | SwinTextSpotter v2 · ICDAR 2015 Strong | H-mean: 80.5 -> 81.8 | Improved |
| add both RC and TLSAM (ablation) | SwinTextSpotter v2 · Total-Text | E2E H-mean: 66.4 -> 71.3 | Improved |
| add local-level attention alignment (ablation) | SwinTextSpotter v2 · Total-Text E2E None | H-mean: 67.9 -> 69.4 | Improved |
| add global-level attention alignment (ablation) | SwinTextSpotter v2 · Total-Text E2E None | H-mean: 67.9 -> 68.7 | Improved |
| increase refinement stages from two to three (ablation) | SwinTextSpotter v2 · Total-Text E2E | H-mean: 65.9 -> 71.3 | Improved |
Later papers that compared with this method
1 later accepted paper reports SwinTextSpotter v2 as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- LRANet++: Low-Rank Approximation Network for Accurate and Efficient Text Spotting (TPAMI 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- SPTS v2 — SPTS v2: Single-Point Scene Text Spotting (TPAMI 2023) · better on 0 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
- SwinTextSpotter: Scene Text Spotting via Better Synergy Between Text Detection and Text Recognition (CVPR 2022 · similarity 0.97)
- SPTS v2: Single-Point Scene Text Spotting (TPAMI 2023 · similarity 0.90)
- Bridging the Gap Between End-to-End and Two-Step Text Spotting (CVPR 2024 · similarity 0.90)
- Towards End-to-End Text Spotting in Natural Scenes (TPAMI 2021 · similarity 0.89)
- GoMatching: A Simple Baseline for Video Text Spotting via Long and Short Term Matching (NeurIPS 2024 · similarity 0.88)
- SemiETS: Integrating Spatial and Content Consistencies for Semi-Supervised End-to-end Text Spotting (CVPR 2025 · 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.
