Table 2.

Haralick’s features description [35].

Haralick’s feature	Equation	Description
Angular second moment	$F_1={\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}P{(i,j)}^2$	measures homogeneity of local gray scale distribution
Contrast	$F_2={\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}{n}^{2}P(i,j)$	$|i-j|=n$, describes quantity of local changes happening in an audio
Correlation	$F_3=\frac{{\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}(\mathit{ij})P(i,j)-{\mu }_x{\mu }_y}{{\sigma }_x{\sigma }_y}$	${\mu }_x$, ${\mu }_y$, ${\sigma }_x$, and ${\sigma }_y$ are mean and standard deviations, value lies between $-1$ to $+1$
Variance	$F_4={\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}{(i-\mu )}^{2}P(i,j)$	$\mu ={\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}\mathit{iP}(i,j)$, measures spread of the signal
Inverse difference moment	$F_5={\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}\frac{P(i,j)}{1+{(i-j)}^2}$	measures local homogeneity
Sum average	$F_6={\sum }_{r=0}^{2N_g-2}{\mathit{rP}}_{x+y}(r)$	measures mean of the gray level sum distribution
Sum variance	$F_7={\sum }_{r=0}^{2N_g-2}{(r-F_6)}^2{P}_{x+y}(r)$	calculates dispersion of the gray level sum distribution
Sum entropy	$F_8=-{\sum }_{i=0}^{2N_g-2}{P}_{x+y}(r)\log (P_{x+y}(r))$	measures disorder related to the gray level sum distribution
Entropy	$F_9=-{\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}P(i,j)\log (P(i,j))$	measure randomness
Difference variance	$F_{10}={\sum }_{r=0}^{N_{g-}1}{(r-{\sum }_{l=0}^{N_{g-}1}{\mathit{lP}}_{|x-y|}(l))}^2{P}_{|x-y|}(r)$	describes heterogeneity
Difference entropy	$F_{11}=-{\sum }_{r=0}^{N_g-1}{P}_{|x-y|}(r)\log (P_{|x-y|}(r))$	measures disorder related to the distribution of gray scale difference
Information measure correlation 1	$F_{12}=\frac{F_9-H_{\mathit{xy}1}}{\mathit{max}\{H_x,H_y\}}$	$H_x$ and $H_y$ are the entropies of $p_x$ and $p_y$ respectively, and $H_{\mathit{xy}1}=-{\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}P(i,j)\log (p_x(i)p_y(j))$
Information measure of correlation 2	$F_{13}=\sqrt{1-\exp (-2(H_{\mathit{xy}2}-F_9))}$	$H_{\mathit{xy}2}=-{\sum }_{i=1}^{N_g}{\sum }_{j=1}^{N_g}{p}_x$ (i) $p_y(j)\log (p_x(i)p_y(j))$
Maximum correlation coefficient	$F_{14}=\sqrt{2^{\mathit{nd}}\mathit{largest}\mathit{eigenvalue}\mathit{of}Q(i,j)}$	$Q(i,j)={\sum }_r\frac{P(i,r)P(j,r)}{p_x(i)p_y(k)}$