From: Comparative analysis of deep learning image detection algorithms
Nº | IoU | Area | maxDets | Average precision | Average recall | ||||
---|---|---|---|---|---|---|---|---|---|
SSD | YOLO | FRCNN | SSD | YOLO | FRCNN | ||||
1 | 0.50:0.95 | All | 100 | 0.247 | 0.337 | 0.716 | 0.232 | 0.279 | 0.782 |
2 | 0.50 | All | 100 | 0.424 | 0.568 | 0.873 | 0.341 | 0.432 | 0.754 |
3 | 0.75 | All | 100 | 0.253 | 0.350 | 0.851 | 0.362 | 0.460 | 0.792 |
4 | 0.50:0.95 | Small | 100/1* | 0.059 | 0.152 | 0.331 | 0.102 | 0.257 | 0.567 |
5 | 0.50:0.95 | Med | 100 | 0.264 | 0.359 | 0.586 | 0.401 | 0.494 | 0.653 |
6 | 0.50:0.95 | Large | 100 | 0.414 | 0.496 | 0.846 | 0.577 | 0.623 | 0.893 |