Scene Text Recognition on ICDAR 2013 (No Lexicon)
94.9Accuracy (No Lexicon)TextScanner+real
Evaluation results
| Method | Links | |
|---|---|---|
| TextScanner+realTraining Data=ST + 90k + real2019.12 | 94.9 | |
| EPTraining Data=ST + 90k2019.12 | 94.4 | |
| SARTraining Data=ST + 90k + real2019.12 | 94 | |
| FANTraining Data=ST + 90k2019.12 | 93.3 | |
| TextScanner+90kTraining Data=ST + 90k2019.12 | 92.9 | |
| TextScanner-preTraining Data=ST2019.12 | 92.7 | |
| TextScanner-mutualTraining Data=ST2019.12 | 92.5 | |
| ASTERTraining Data=ST + 90k2019.12 | 91.8 | |
| CA-FCNTraining Data=ST2019.12 | 91.5 | |
| Jaderberg et al. (2016)Training Data=90k2019.12 | 90.8 | |
| R2AMTraining Data=90k2019.12 | 90 | |
| RARETraining Data=90k2019.12 | 88.6 | |
| CRNNTraining Data=90k2019.12 | 86.7 | |
| Jaderberg et al. (2014a)Training Data=90k2019.12 | 81.8 |