Table 2 Kappa statistics of agreement of measurement results on success/failure of tooth segmentation.

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

Segmentation type

 

First

Kappa

p value

Success

Failure

Total

LS

Second

Success

271 (100.0%)

0 (0.0%)

271 (96.8%)

1.000

 < 0.0001

Failure

0 (0.0%)

9 (100.0%)

9 (3.2%)

Total

271 (100.0%)

9 (100.0%)

280 (100.0%)

DS

Second

Success

244 (99.6%)

0 (0.0%)

244 (87.1%)

0.984

 < 0.0001

Failure

1 (0.4%)

35 (100.0%)

36 (12.9%)

Total

245 (100.0%)

35 (100.0%)

280 (100.0%)

AS

Second

Success

270 (99.3%)

0 (0.0%)

270 (96.4%)

0.885

 < 0.0001

Failure

2 (0.7%)

8 (100.0%)

10 (3.6%)

Total

272 (100.0%)

8 (100.0%)

280 (100.0%)

  1. Success and failure of segmentation, presented as nominal variables using Cohen's kappa, indicated very high evaluation reproducibility from 0.885 to 1.000.
  2. Kappa values were interpreted as follows: poor,  < 0.0; slight, 0.0 to 0.2; fair, 0.2 to 0.4; moderate, 0.4 to 0.6; substantial, 0.6 to 0.8, and almost perfect, 0.8 to 1.0.