Skip to main content
. 2022 Sep 28;13:993244. doi: 10.3389/fpls.2022.993244

TABLE 2.

Test performance comparison of different models.

Methods RMSE MAE Recall mAP@.0.5 Map@.0.5:0.95
YOLOv5s 53.23 41.24 0.887 0.949 0.526
YOLOv5m 51.56 40.83 0.894 0.949 0.522
YOLOv5l 49.71 38.87 0.888 0.947 0.525
YOLOv5x 44.51 33.62 0.913 0.950 0.541
Improved YOLOv5s 43.94 34.36 0.911 0.951 0.545
Faster R-CNN 94.57 87.10 0.819 0.862 0.355