Table 6 Improved YOLOv8s and YOLOv8s experiments performance.

From: BD-YOLOv8s: enhancing bridge defect detection with multidimensional attention and precision reconstruction

algorithms

ODConv (1)

C2f_cbam (2)

CARAFE (3)

mAP@0.5 (%)

mAP@0.5:0.9 (%)

FPS/HZ

F1

Params (M)

GFLOPs

YOLOv8s

   

80.9

50.3

105

80

11.13

28.4

 + (1)

√

  

86.6

53.6

106

85

11.19

27.2

 + (2)

 

√

 

85.7

52.8

103

84

11.2

28.9

 + (3)

  

√

80.9

51.3

94

79

11.6

30.3

 + (1) + (2)

√

√

 

86.1

53.8

99

84

11.26

27.6

 + (2) + (3)

 

√

√

82.2

53.6

89

80

11.67

30.8

 + (1) + (3)

√

 

√

82.4

52

92

81

11.66

29.0

BD-YOLOv8s

√

√

√

86.2

56

84

89

11.73

29.5

Back to article page