Table 6 Statistical analysis of main effects and first-order interactions affecting measurement error using GLMM.

From: Accuracy and efficiency of automatic tooth segmentation in digital dental models using deep learning

Variable

DF

F

pa value

Post hoc testb

Interceptc

1:4723

319.03

 < 0.001

 

Tooth group

3:4723

296.67

 < 0.001

Molar (0.102) > incisal (−0.098), Canine (−0.110) > premolar (−0.149)

Software

2:4723

70.00

 < 0.001

DS (−0.019) > AS (−0.052) > LS (−0.121)

MD width/CCH

1:4723

32.92

 < 0.001

MD width (−0.043) > CCH (−0.084)

Tooth group * Software

6:4723

24.70

 < 0.001

 

Tooth group * MD width/CCH

3:4723

201.37

 < 0.001

 

Software * MD width/CCH

2:4723

428.81

 < 0.001

 
  1. The tooth size errors were statistically different (p < 0.001) depending on the software used, and the post hoc test showed that DS (−0.019) > AS (−0.052) > LS (−0.121).
  2. DF degrees of freedom.
  3. F F value.
  4. ap values were derived from a generalised linear mixed model.
  5. bCategory (estimated mean) was presented for Bonferroni’s corrected post hoc test.
  6. cIntercept represents the mean value of the response variable when all predictor variables in the model are zero.
Back to article page