Research Article
Aircraft Detection for Remote Sensing Images Based on Deep Convolutional Neural Networks
Table 2
The comparison of performance indicators for the aircraft.
| ā | TP | FP | FN | Precision (%) | Recall (%) | F1-score (%) | IOU (%) | AP (%) | FPS |
| YOLOv3 | 936 | 99 | 116 | 90.43 | 88.97 | 89.70 | 67.42 | 89.24 | 30 | MSDN | 1000 | 30 | 52 | 97.09 | 95.06 | 96.06 | 75.76 | 90.64 | 28 | MSRDN-F | 998 | 18 | 54 | 98.23 | 94.87 | 96.52 | 78.91 | 90.67 | 25 | MSRDN-M | 1002 | 18 | 50 | 98.24 | 95.25 | 96.72 | 79.06 | 90.66 | 25 | MSRDN-B | 995 | 25 | 57 | 97.55 | 94.58 | 96.04 | 78.35 | 90.66 | 25 |
|
|