Table 4 Results of character recognition.
From: A large-scale dataset for Chinese historical document recognition and analysis
Training set | MTHv2 | M5HisDoc | CASIA-AHCDB | HisDoc1B |
|---|---|---|---|---|
MTHv2 | 75.27 | 34.59 | 55.87 | 26.89 |
M5HisDoc | 82.27 | 66.49 | 79.35 | 52.28 |
CASIA-AHCDB | 71.92 | 38.32 | 95.21 | 28.84 |
HisDoc1B (Ours) | 85.83 | 79.01 | 92.13 | 94.03 |
