relative mutual information Uxy

Uxy calculates the relative mutual information (and other quantities) derived from grey-value cooccurence matrices

The function Uxy and the two subfunctions that are called by Uxy calculate the relative mutual information associated to grey-value cooccurence matrices. The function simply implements the equations given in the cited papers by Nowosad and Stepinski (2019) and by Haralick (1973). Please consult the Nowosad and Stepinski paper for the definitions of the output (Hygx stands for H(y|x), the rest should be obvious)

The input Y may be a simple matrix (grey value image) or an array of images. The input thr is a structure (or, if several images are analysed at once, an array of structures). thr.values (or thr(k).values) contains threshold values that are chosen to segment the image(s) into separate grey value levels. The function greyThreshO, which again calls the function multithresh (Otsu's method) from the image processing toolbox, may be used for this. The function is available at the Matlab Central as well, as part of my submission deltaEtAl (see and download there). However, the users may supply their own function to provide the structure thr, in that case, the image processing toolbox is not required. If the input iNN is set to 1, only the four next neighbours are considered to calculate the cooccurence matrices; if it is set to 0, all eight next neighbours will be taken into account. The input iguck should be integer. If it is different from zero, the outputs of the function Uxy are visualised in figure(iguck). If iguck==0, there is no graphic output.