From: Chinese text-line detection from web videos with fully convolutional networks
Method | Dataset I | Dataset II | Dataset III | ||||||
---|---|---|---|---|---|---|---|---|---|
 | Precision | Recall | F1 | Precision | Recall | F1 | Precision | Recall | F1 |
Yin et al. | 79.82 | 74.28 | 76.95 | 80.15 | 75.38 | 77.69 | 80.02 | 74.37 | 77.09 |
Ours+ResNet-50 | 82.33 | 84.85 | 83.57 | 82.39 | 85.62 | 83.97 | 82.11 | 81.88 | 81.99 |
Ours+PVANET | 83.02 | 84.97 | 83.98 | 80.63 | 86.20 | 83.32 | 81.98 | 85.65 | 83.74 |