Skip to main content

Table 5 Summary of the key performance metrics for the CNN models

From: Impacted lower third molar classification and difficulty index assessment: comparisons among dental students, general practitioners and deep learning model assistance

Model

Accuracy

AUC

Processing time (seconds)

Angulation

ā€ƒRegNetY032

0.82

0.98

3.08 ± 0.10

ā€ƒDenseNet201

0.76

0.96

5.57 ± 0.07

ā€ƒXception

0.72

0.94

1.77 ± 0.07

ā€ƒInceptionV3

0.66

0.89

2.52 ± 0.08

ā€ƒResNetRS101

0.28

0.54

5.14 ± 0.10

ā€ƒInceptionResNetV2

0.88

0.97

6.01 ± 0.04

Class

ā€ƒRegNetY032

0.68

0.85

3.08 ± 0.10

ā€ƒDenseNet201

0.74

0.86

5.57 ± 0.07

ā€ƒXception

0.78

0.87

1.77 ± 0.07

ā€ƒInceptionV3

0.60

0.74

2.52 ± 0.08

ā€ƒResNetRS101

0.38

0.61

5.14 ± 0.10

ā€ƒInceptionResNetV2

0.62

0.83

6.01 ± 0.04

Position

ā€ƒRegNetY032

0.76

0.87

3.08 ± 0.10

ā€ƒDenseNet201

0.60

0.85

5.57 ± 0.07

ā€ƒXception

0.90

0.96

1.77 ± 0.07

ā€ƒInceptionV3

0.80

0.92

2.52 ± 0.08

ā€ƒResNetRS101

0.60

0.83

5.14 ± 0.10

ā€ƒInceptionResNetV2

0.92

0.96

6.01 ± 0.04