how to determine multilevel thresholds for segmentation using real coded genetic agorithm

function fitR = Kapur(m,level,xR,PI) %Metodo de Entropia de Kapur fitR = zeros(1,m); for j = 1: m PI0 = PI(1:xR(j,1)); % probabilidad de la primer clase ind = PI0 == 0; ind = ind .* eps; PI0 = PI0 + ind; clear ind w0 = sum(PI0); %w0 de la primer clase H0 = -sum((PI0/w0).*(log2(PI0/w0))); fitR(j) = fitR(j) + H0;

    for jl = 2: level
        PI0 = PI(xR(j,jl-1)+1:xR(j,jl)); % probabilidad de la primer clase
        ind = PI0 == 0;
        ind = ind .* eps;
        PI0 = PI0 + ind;
        clear ind
        w0 =  sum(PI0); %w0 de la primer clase
        H0 = -sum((PI0/w0).*(log2(PI0/w0)));
        fitR(j) = fitR(j) + H0;
    end  
end

In this code how to determine the level, xR, PI, and m