Skip to main content

Table 1 Prediction errors in 4 different settings (unit: mm)

From: Automatic craniomaxillofacial landmarks detection in CT images of individuals with dentomaxillofacial deformities by a two-stage deep learning model

 

R-ZR

L-ZR

FR

NR

TR

MER

R-MAR

L-MAR

SR

Overall

V-Net without

Region division

         

2.40

± 1.08

3D UX-Net without

Region division

         

2.34

± 1.01

V-Net with

Region division

1.82

± 0.84

1.92

± 0.93

1.61

± 0.80

1.99

± 1.08

1.89

± 0.94

2.16

± 1.01

2.18

± 0.99

1.89

± 0.84

1.26

± 0.52

1.90

± 0.93

3D UX-Net with

Region division

1.74

± 0.82

2.02

± 0.98

1.58

± 0.77

1.93

± 0.96

1.74

± 0.88

2.23

± 1.09

1.90

± 0.96

1.97

± 0.89

1.22

± 0.44

1.81

± 0.89