Table 2 Non parametric MANCOVA on MAD of the models obtained through different thresholds from the reference model, by different machines, thresholds, and superimposition statuses. Three factors and their interactions were analyzed having “patient” as a covariate: machine (random factor; 4 machines), threshold (random factor; 6 thresholds) and superimposition status (fixed factor; before and after superimposition). Machine and threshold were crossed factors, whereas superimposition was nested in threshold factor. *p < 0.05.

From: The effect of threshold level on bone segmentation of cranial base structures from CT and CBCT images

Distance factor

df

F

P(perm)

Covariate

1

206.3

0.001*

Machine

3

8.6

0.002*

Threshold

5

40.8

0.001*

Superimposition (Threshold)

6

3.4

0.024*

Machine × Threshold

15

3.3

0.001*

Machine × Superimposition (Threshold)

18

6.1

0.001*

Residual

431

  

Total

479

  

Comparisona

P

  

A vs. B

0.005*

  

A vs. C

0.005*

  

A vs. D

0.091

  

B vs. C

0.060

  

B vs. D

0.457

  

C vs. D

0.645

  
  1. aTests among levels of the factor Machine. A, B, C, and D correspond to the four radiographic machines tested in the study.
Back to article page