Table 3.

GLCM texture parameters [30] where (i): reference cell in the co-occurrence matrix; (j): neighbor cell in the co-occurrence matrix; Pij: GLCM expressed as a probability of having a pair of cell values with a specific value in a specific spatial relationship, µ: mean and σ: standard deviation within GLCM in the specified window size.

Texture Measures	Formula	Explanation
Entropy	∑ − ln(Pij) Pij	Smaller Pij leads to higher entropy value
Contrast	∑Pij (i − j)2	Express difference as an exponential function
Variance	∑Pij (i − µi)2	Describe the variance of GLCM values
Correlation	∑Pij [(i − µi) (j − µi)/σ2]	Describe the correlation of GLCM values
Mean	∑Pij/N	Describe the mean of the GLCM values
Homogeneity	∑Pij/(1 + (i − j))2	Express difference as an inversed exponential function
Dissymmetry	∑(∑(|i − j| Pij))	Express difference as a linear function
Second moment	∑(∑(Pij)2)	Return the max value when all pixels are identical