modified Minimum Cross Entropy threshold selection

% MinCEP is a function for thresholding using Minimum Cross Entropy
% threshold selection of non-blank space of Image.
%
% [ILow, IHigh, threshold] = minCEP(I)
% Inputs: I 2D grayscale image
% minIntensity minimum intensity
% maxIntensity maximum intensity
%
% Outputs: ILow image with gray level bellow threshold
% Outputs: IHigh image with gray level upper threshold
% Outputs: threshold the threshold choosen by MinCE
%
% Description: This code implements the paper: "Minimum cross-entropy
% threshold selection" by Brink and Pendock
% This implemetation is modified to calculate threshold of non-blank space
% of Image as well as get minimum and maximum intensity of image.
%
% Example:
% img = imread('tire.tif');
% [ILow, IHigh, T] = minCEP(img);
% subplot(2,2,1), imshow(img);
% subplot(2,2,2), imshow(IHigh);
% [~, IHigh, T1] = minCEP(img, T);
% subplot(2,2,3), imshow(IHigh);
% [~, IHigh, T2] = minCEP(img, T2);
% subplot(2,2,4), imshow(IHigh);
%
%
% Coded by: Hoosein Dehghan (hd.dehghan@gmail.com)
%------------------------------------------------------------------------