Table 2 Comparison of PCE-YOLO and other algorithms of different scales on various datasets

From: An efficient algorithm for pedestrian fall detection in various image degradation scenarios based on YOLOv8n

Datastes

Fall

Low-Light_fall

Occlusion_fall

Fuzzy_fall

Algorithms

Metrics

PCE-YOLO

Precision/%

92.46

90.3

88.72

87.36

Recall/%

85.97

83.23

78.13

74.53

mAP0.5/%

92.94

91.47

88.14

87.67

mAP0.5:0.95/%

65.18

61.01

58.63

60.1

YOLOv8n

Precision/%

89.32

88.8

83.18

81.01

Recall/%

80.36

77.5

77.99

72.18

mAP0.5/%

89.63

87.07

83.43

81.96

mAP0.5:0.95/%

61.79

56.81

53.18

52.55

YOLOv8s

Precision/%

92.15

90.1

86.34

84.02

Recall/%

83.26

81.22

79.03

78.4

mAP0.5/%

91.4

89.9

85.11

84.89

mAP0.5:0.95/%

62.48

58.42

56.25

55.18

YOLOv9s

Precision/%

92.71

90.47

86.79

85.2

Recall/%

83.88

82.05

79.25

77.53

mAP0.5/%

91.84

90.88

85.64

85.71

mAP0.5:0.95/%

63.32

59.47

56.93

57.86

YOLOv10s

Precision/%

93.5

90.62

87.03

86.37

Recall/%

84.1

82.5

79.41

77.79

mAP0.5/%

92.56

91.3

86.17

86.6

mAP0.5:0.95/%

64.4

60.62

57.06

58.42