Scene Text Recognition on ICDAR 50-word lexicon 2003 (test)
99.2AccuracyDICT-IC03-Full
Evaluation Results
| Method | Links | |
|---|---|---|
| DICT-IC03-FullLanguage Model=Tightly constrained (IC03-Full)2014.06 | 99.2 | |
| DICT+2-90kLexicon=90k-word, Model Size=490M parameters2014.06 | 98.7 | |
| CHAR+2Language Model=None2014.06 | 96.7 | |
| NGRAM+2-SVMClassifier=Linear SVM2014.06 | 96.5 | |
| Jaderberg2014.06 | 96.2 | |
| DICT+2-90kLexicon=90k-word2014.06 | 93.3 | |
| Alsharif2014.06 | 93.1 | |
| Wang & Wu2014.06 | 90 | |
| Goel2014.06 | 89.7 | |
| CHAR+2Language Model=None2014.06 | 89.5 | |
| Yao2014.06 | 88.5 | |
| Alsharif2014.06 | 85.1 | |
| Novikova2014.06 | 82.8 | |
| Mishra2014.06 | 81.8 | |
| Wang2014.06 | 76 | |
| Baseline ABBYY2014.06 | 56 |