Table 3 Text detection results on three experimental scenarios using different models
From: Detection and recognition of cursive text from video frames
Models | Scenario | Precision | Recall | F-Measure |
|---|---|---|---|---|
SSD | S-I | 79.7 | 66.9 | 72.7 |
| Â | S-II | 75.3 | 56.1 | 64.3 |
| Â | S-III | 80.3 | 70.5 | 75.1 |
Faster R-CNN | S-I | 83.1 | 84.9 | 84.0 |
| Â | S-II | 74.7 | 67.6 | 71.0 |
| Â | S-III | 86.9 | 89.8 | 88.3 |
East | S-I | 78.2 | 45.3 | 57.3 |
| Â | S-II | 70.4 | 47.7 | 56.9 |
| Â | S-III | 77.3 | 47.9 | 59.1 |