Table 1.

The best performing two- and three-feature combinations as judged by the AUC and the confidence interval

Two-features combination	AUC	Confidence interval
Kurtosis, coarseness	0.92	[0.90, 0.95]
NESTD, coarseness	0.90	[0.89, 0.97]
Kurtosis, NESTD	0.89	[0.84, 0.91]
Kurtosis, skewness	0.88	[0.84, 0.92]
Two features combination	AUC	Confidence interval
Kurtosis, coarseness, NESTD	0.94	[0.91, 0.95]
Skewness, contrast, IDF	0.93	[0.91, 0.95]
Kurtosis, NESTD, skewness	0.82	[0.90, 0.95]
Kurtosis, skewness, mean	0.88	[0.85, 0.91]