Skip to main content

Table 5 Comparison of adding different attention mechanisms in YOLOv8

From: Melon ripeness detection by an improved object detection algorithm for resource constrained environments

Methods

P(%)

R(%)

mAP(%)

Parameters(M)

FLOPs(G)

FPS(GPU)

GAM

94.5

93.4

97.1

3.70

9.6

81.76

PolarizedSelf

92.2

95.0

97.2

3.22

8.6

82.30

NAM

93.6

93.3

96.7

3.01

8.2

86.85

SE

96.1

91.6

96.9

3.01

8.2

83.45

Shuffle

93.1

93.7

97.1

3.01

8.2

82.44

SimAM

95.3

90.9

97.0

3.01

8.2

89.13

CBAM

94.3

93.4

97.2

3.11

8.3

81.97

CA

93.1

94.3

97.3

3.03

8.2

84.96