| Baidu IDL | 100.00% | 100.00% | 100.00% |
| Yunos_Robot1.0 | 100.00% | 100.00% | 100.00% |
| Baidu VIS | 100.00% | 100.00% | 100.00% |
| HoText_v1 | 100.00% | 100.00% | 100.00% |
| Baidu VIS v2 | 100.00% | 100.00% | 100.00% |
| Tencent Youtu | 100.00% | 100.00% | 100.00% |
| Baidu VIS v2+ | 100.00% | 100.00% | 100.00% |
| Sogou_MM | 100.00% | 100.00% | 100.00% |
| Alibaba-PAI | 100.00% | 100.00% | 100.00% |
| gts-swint-cocoonly-s4-3_1 | 100.00% | 100.00% | 100.00% |
| Baidu IDL v3 | 100.00% | 87.50% | 93.33% |
| TextProposals + DictNet | 85.71% | 100.00% | 92.31% |
| SRC-B-TextProcessingLab | 85.71% | 100.00% | 92.31% |
| SRC-B-MachineLearningLab | 85.71% | 100.00% | 92.31% |
| SRC-B-MachineLearningLab | 71.43% | 83.33% | 76.92% |
| Sensetime V1 | 71.43% | 83.33% | 76.92% |
| FOTS | 71.43% | 83.33% | 76.92% |
| e2e-mlt | 71.43% | 71.43% | 71.43% |
| NVTextSpotter | 71.43% | 71.43% | 71.43% |
| e2e-mlt + bug fix | 71.43% | 62.50% | 66.67% |
| Char-level PTSD on SynthText | 57.14% | 66.67% | 61.54% |
 Baseline OpenCV3.0 + Tesseract | 28.57% | 100.00% | 44.44% |
| Deep2Text-MO | 42.86% | 42.86% | 42.86% |
| Megvii-Image++ | 42.86% | 37.50% | 40.00% |
| M_FOTS_Orig | 42.86% | 37.50% | 40.00% |
| Beam search CUNI +S (decoding for TextSpotter - no postprocessing) | 28.57% | 40.00% | 33.33% |
| Baseline-TextSpotter | 0.00% | 0.00% | 0.00% |
| Beam search CUNI (decoding for TextSpotter - spell checked) | 0.00% | 0.00% | 0.00% |