Skip to main content

Table 2 Segmentation accuracy according to parameters assigned based on P1 and P2

From: Accuracy of advanced deep learning with tensorflow and keras for classifying teeth developmental stages in digital panoramic imaging

  

P1

P2

Radius length, in pixels

Number of radial lines

Output state

Time (s)

Dice Coefficient (F1 Score)

Jaccard similarity

Output state

Time (s)

Dice Coefficient (F1 score)

Jaccard similarity

80

800

*

1.64

0.85

0.76

*

1.74

0.81

0.73

 

1000

*

1.75

0.87

0.78

*

1.83

0.83

0.74

 

1200

***

2.36

0.88

0.75

*

1.97

0.85

0.78

100

800

*

2.01

0.85

0.74

*

2.13

0.86

0.75

 

1000

**

2.22

0.86

0.77

**

2.29

0.91

0.86

 

1200

**

2.42

0.96

0.90

**

2.31

0.94

0.88

120

800

**

2.54

0.94

0.85

*

2.68

0.93

0.84

 

1000

*

2.92

0.89

0.73

**

2.77

0.80

0.71

 

1200

**

3.26

0.92

0.83

*

3.64

0.92

0.83

140

800

**

2.97

0.92

0.81

*

3.21

0.93

0.84

 

1000

**

3.03

0.90

0.79

***

3.98

0.89

0.81

 

1200

***

3.93

0.85

0.72

***

4.13

0.84

0.76

  1. *Under-segmented, **well-segmented, ***over-segmented
Back to article page

BMC Medical Imaging

ISSN: 1471-2342

Contact us