Skip to main content
. 2024 Jan 15;14:1323301. doi: 10.3389/fpls.2023.1323301

Table 3.

Comparison with different target detection models.

Methods mAP@0.5
(%)
mAP@ [0.5:0.95]
(%)
Precision
(%)
Recall
(%)
SSD 66.2 46.9 96.8 36.9
YOLOV3 65.1 33.3 64.3 63.4
YOLOV4 52.4 18.6 87.9 25.1
YOLOV5 67.8 34.9 67.3 66.1
MGA-YOLOV5 69.0 34.2 74.4 64.3
BTC-YOLOV5 72.0 39.8 70.7 67.9
YOLOV5-CBAM-C3TR 73.4 40.9 70.9 69.5

The performance of the indicators is best shown in bold.