Table 2 Average evaluation segmentation performance (%) over five cross-validation per class for U-Net and ResU-Net using the testing MRI datasets.

From: A hybrid segmentation and classification CAD framework for automated myocardial infarction prediction from MRI images

Tissue Type

Classes

U-Net

ResU-Net

ACC

SEN

Pre

F1-score

MIoU

ACC

SEN

Pre

F1-score

MIoU

Normal Regions

BG

98.74

97.25

98.02

97.63

95.05

99.78

99.04

99.23

99.13

98.59

LV

89.23

86.45

88.03

87.23

85.05

95.89

89.52

92.73

91.12

91.86

Myo

86.72

83.76

82.58

83.17

80.49

91.28

88.67

88.12

88.39

87.84

Abnormal Regions

MI

78.24

75.53

74.87

75.21

73.25

85.15

81.21

78.98

80.08

77.99

MVO

63.02

61.92

63.47

62.69

58.63

70.31

67.78

68.25

68.01

64.89

Avg.

All

83.19

80.98

81.39

81.19

78.49

88.48

85.24

85.46

85.35

84.23