Table 4 The 2-fold comparison results.

From: Optimizing deep belief network for concrete crack detection via a modified design of ideal gas molecular dynamics

Method

Accuracy (%)

Specificity (%)

Fl Score (%)

Precision (%)

Recall (%)

FCN9

78.283

73.428

64.735

76.081

76.983

YOLO10

75.224

80.195

76.692

82.451

80.926

CrackSegNet11

74.294

68.524

68.486

79.469

85.811

CNN12

77.325

77.278

77.048

82.46

89.326

CedNet13

74.018

72.938

76.93

84.965

89.145

DBN/MIGMM

87.601

91.457

78.495

89.667

90.062

Back to article page