Research map / Video / Video-language understanding
A Bilingual, OpenWorld Video Text Dataset and End-to-end Video Text Spotter with Transformer
Weijia Wu, Debing Zhang, Yuanqiang Cai, Sibo Wang, Jiahong Li, Zhuang Li, Yejun Tang, Hong Zhou
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 |
|---|---|---|---|
| Use TransVTSpotter instead of the strongest listed baseline detector | BOVText text detection · BOVText | F-score: 80.8 -> 81.7 | Improved |
| Use TransVTSpotter instead of DB + CRNN | BOVText text tracking · BOVText | ID_F1: 48.3 -> 64.7 | Improved |
| Use TransVTSpotter instead of DB + RARE | BOVText end-to-end spotting · BOVText | ID_F1: 34.2 -> 40.8 | Improved |
| Train TransVTSpotter on BOVText rather than LSVTD | TransVTSpotter · BOVText NewsReport scenario | ID_F1: 42.1 -> 70.4 | Improved |
| Add angle prediction (ablation) | TransVTSpotter · ICDAR2015(video) tracking | ID_F1: 36.1 -> 57.3 | Improved |
| Use cosine loss rather than L1 loss for angle prediction (ablation) | TransVTSpotter · ICDAR2015 text detection | F-score: 83.2 -> 84.2 | Improved |
| Add tracking query to the object query (ablation) | TransVTSpotter · ICDAR2015(video) tracking | ID_F1: 54.2 -> 57.3 | Improved |
| Increase input short side from 480 to 800 (ablation) | TransVTSpotter · ICDAR2015(video) tracking | ID_F1: 52.6 -> 57.3 | Improved |
| Use TransVTSpotter instead of Free | ICDAR2015(video) tracking · ICDAR2015(video) | MOTA: 43.2 -> 44.1 | Improved |
| Use TransVTSpotter instead of Free | ICDAR2015(video) spotting · ICDAR2015(video) | ID_F1: 61.8 -> 61.5 | Hurt |
| Use TransVTSpotter rather than Wei et al. | Video text tracking · Minetto(video) | MOTA: 83.5 -> 84.1 | Improved |
Closest papers by meaning
- OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation (ICLR 2025 · similarity 0.88)
- GoMatching: A Simple Baseline for Video Text Spotting via Long and Short Term Matching (NeurIPS 2024 · similarity 0.86)
- Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers (CVPR 2024 · similarity 0.86)
- TOPA: Extending Large Language Models for Video Understanding via Text-Only Pre-Alignment (NeurIPS 2024 · similarity 0.86)
- Mimir: Improving Video Diffusion Models for Precise Text Understanding (CVPR 2025 · similarity 0.86)
- Vript: A Video Is Worth Thousands of Words (NeurIPS 2024 · similarity 0.86)
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.
