Table 1.

Summary of Different Texture Feature Categories

Category	Texture Features	Number
Histogram	mean, variance, skewness, kurtosis, 1-% percentile, 10-% percentile, 50-% percentile, 90-% percentile, and 99-% percentile	9
Gradient	mean, variance, skewness, kurtosis, and percentage of pixels with nonzero gradient	5
Run-length matrix*	run length nonuniformity, gray level nonuniformity, long run emphasis, short run emphasis, and fraction of image in runs	20
Co-occurrence matrix†	angular second moment, contrast, correlation, sum of squares, inverse difference moment, sum average, sum variance, sum entropy, entropy, difference variance, and difference entropy	220
Wavelet analysis	teta1, teta2, teta3, teta4, and sigma	5
Autoregressive model‡	WavEn	20

^*Each feature in this category is computed for four directions for each tumor (d = 0°, 45°, 90°, and 135°).

^†Each feature in this category is computed for four directions and five between-pixel distances for each tumor (d = 0°, 45°, 90°, and 135°; θ = 1, 2, 3, 4, and 5).

^‡This feature is computed at five scales within four frequency bands; thus, the total number of features calculated for this category is twenty.