Table 9 Approach (2): execution computational cost for all AI classification models over all scenarios.

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

Scenario

AI Model

No. of Trainable Parameters (Million)

Training Time/Epoch

(msec)

Testing Time/Image (sec)

Frame Per Second (FPS)

ACC

(%)

Scenario 1

ResNet50

50.79

24

0.64

1.54

94.44

VGG16

20.89

28

0.56

1.79

93.52

ResNet50-V2

51.64

25

0.74

1.35

90.74

Xception

10.50

63

2.88

0.35

88.89

InceptionResNetV2

18.48

308

1.49

0.67

87.96

Scenario 2

Ensemble 1

70.17

50

0.97

1.03

94.44

Ensemble 2

76.21

67

1.26

0.79

95.37

Scenario 3

ResNet + ViT

88.31

57

1.53

0.65

95.37

Ensemble 1 + ViT

89.29

191

2.29

0.44

97.22

Ensemble 2 + ViT

89.96

192

2.31

0.43

98.15

Back to article page