Research map / Vision-language and multimodal / CLIP and zero-shot
Language Matters: A Weakly Supervised Vision-Language Pre-training Approach for Scene Text Detection and Spotting
Chuhui Xue, Wenqing Zhang, Yu Hao, Shijian Lu, Philip H. S. Torr, Song 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 oCLIP weak pretraining on LSVT | DB · ICDAR2019-LSVT detection | F-score: 73.4 -> 75.8 | Improved |
| add oCLIP weak pretraining on LSVT | PSENet · ICDAR2019-LSVT detection | F-score: 74.6 -> 77.1 | Improved |
| add oCLIP weak pretraining on LSVT | Mask TextSpotter-V3 · ICDAR2019-LSVT spotting | E2E F-score: 27.7 -> 32.5 | Improved |
| pretrain with oCLIP instead of no pretraining | PSENet · Total-Text | F-score: 78.3 -> 85.5 | Improved |
| use oCLIP pretraining instead of SynthText-trained baseline | PSENet · Total-Text | F-score: 82.6 -> 85.5 | Improved |
| use oCLIP instead of STKM pretraining on the same data | PSENet · Total-Text | F-score: 82.2 -> 85.5 | Improved |
| use oCLIP instead of STKM pretraining on the same data | PSENet · CTW1500 | F-score: 81.5 -> 82.8 | Improved |
| pretrain on 40M automatically extracted web images instead of SynthText | PSENet+oCLIP · Total-Text | F-score: 85.5 -> 87.0 | Improved |
| annotate 100% instead of 25% of text instances in pretraining (ablation) | PSENet+oCLIP · Total-Text | F-score: 84.8 -> 85.5 | No clear effect |
| add oCLIP pretraining | DB-RN50 · Total-Text | F-score: 78.2 -> 84.1 | Improved |
| add oCLIP pretraining | DB-RN50 · ICDAR2015 | F-score: 80.9 -> 85.4 | Improved |
| add oCLIP pretraining | Mask TextSpotter-V3 · ICDAR2015 spotting (G lexicon) | E2E accuracy: 74.2 -> 74.3 | No clear effect |
Closest papers by meaning
- ODM: A Text-Image Further Alignment Pre-training Approach for Scene Text Detection and Spotting (CVPR 2024 · similarity 0.90)
- InstructOCR: Instruction Boosting Scene Text Spotting (AAAI 2025 · similarity 0.89)
- Image as a Language: Revisiting Scene Text Recognition via Balanced, Unified and Synchronized Vision-Language Reasoning Network (AAAI 2024 · similarity 0.89)
- OTE: Exploring Accurate Scene Text Recognition Using One Token (CVPR 2024 · similarity 0.88)
- CLIP is Almost All You Need: Towards Parameter-Efficient Scene Text Retrieval without OCR (CVPR 2025 · similarity 0.88)
- SuperCLIP: CLIP with Simple Classification Supervision (NeurIPS 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.
