Skip to main content

Advances, Systems and Applications

Table 3 Comparison of evaluation indicators of different models

From: Small object Lentinula Edodes logs contamination detection method based on improved YOLOv7 in edge-cloud computing

Model

mAP@0.5

Recall

mAP@0.5:0.9

FPS

ResNeXt-50 (32 × 4d)

93.32%

91.12%

87.13%

17

MobilenetV3-YOLOv4

92.16%

89.11%

84.77%

30

Ghost-YOLOv4

93.04%

91.15%

86.56%

36

YOLOv5l

94.79%

92.68%

87.51%

67

YOLOv6m

95.82%

93.29%

91.32%

48

YOLOv7

96.31%

92.80%

91.66%

56

SRW-YOLO

98.62%

96.43%

92.39%

62