MatlabCompat.jl - MATLAB/Octave functions library for Julia

![MatlabCompat.jl](img/logo.png "MatlabCompat.jl") ### MATLAB/Octave functions library for Julia, fork us on on [GitHub](http://github.com/MatlabCompat/MatlabCompat.jl) ###[__Home__](index.html) : [__Help__](help.html) : [__Contribute__](contribute.html) ## Installation and Getting started * In your Julia console invoke: ``` Pkg.add("MatlabCompat") ``` * Restart julia kernel if you are using IJulia. ### Dependencies We use following Julia pacakages: Tk Images ImageView FixedPointNumbers MAT Color BinDeps ##Illustrative Examples <a name="example1"></a> To illustrate how MatlabCompat.jl can be used to convert MATLAB/Octave code to Julia we have created a simplistic image analysis script in MATLAB/Octave we called janus.m: ``` tic() % read the remote image img = imread('http://matlabcompat.github.io/img/example.tif'); % compute graysacale threshold using Otsu algorithm threshold = graythresh(img); % create a binary image based on the grayscale image imgbw = im2bw(img, threshold); % display the resulting binary image imshow(imgbw); % lable each connected object in the image labeledbw = bwlabel(imgbw, 4); % count cells numberOfCells = max(reshape(labeledbw, 1,numel(labeledbw))); % display the number of objects disp(strcat('number of objects:', num2str(numberOfCells))); imshow(label2rgb(labeledbw,'jet',[0 0 0],'shuffle')); toc() ``` by invoking rosetta(janus.m, janus.jl) function once can convert jamus.m to julia (janus.jl): ``` using MatlabCompat importall MatlabCompat.ImageTools importall MatlabCompat.MathTools tic() # read the remote image img = imread("http://matlabcompat.github.io/img/example.tif"); # compute graysacale threshold using Otsu algorithm threshold = graythresh(img); # create a binary image based on the grayscale image imgbw = im2bw(img, threshold); # display the resulting binary image imshow(imgbw); # lable each connected object in the image labeledbw = bwlabel(imgbw, 4); # count cells numberOfCells = max(reshape(labeledbw, 1,numel(labeledbw))); disp(strcat("number of objects:", num2str(numberOfCells))); # display the number of objects imshow(label2rgb(labeledbw,"jet",[1 1 1],"shuffle")); toc() ``` <a name="example2"></a> Similar approach can be used for the slightly modified janus2.m: ``` % read the remote images imgTreated = imread('http://matlabcompat.github.io/img/example_Treated.tif'); imgNonTreated = imread('http://matlabcompat.github.io/img/example_NonTreated.tif'); % compute grayscale threshold of the non treated image using Otsu algorithm threshold = graythresh(imgNonTreated); % create a binary image based on the grayscale images imgBWTreated = im2bw(imgTreated, threshold); imgBWNonTreated = im2bw(imgNonTreated, threshold); % count foreground pixels for each image and output it % in console foregroundPixelCountTreated = sum(reshape(imgBWTreated, 1,numel(imgBWTreated))); foregroundPixelCountNonTreated = sum(reshape(imgBWNonTreated, 1,numel(imgBWNonTreated))); display(strcat('Foreground pixel count for Treated specimen: ',num2str(foregroundPixelCountTreated))) display(strcat('Foreground pixel count for Non-Treated specimen: ',num2str(foregroundPixelCountNonTreated))) ``` Which can be converted to Julia language automatically using rosetta() function yielding janus2.jl: ``` using MatlabCompat importall MatlabCompat.ImageTools importall MatlabCompat.MathTools # read the remote images imgTreated = imread("http://matlabcompat.github.io/img/example_Treated.tif"); imgNonTreated = imread("http://matlabcompat.github.io/img/example_NonTreated.tif"); # compute grayscale threshold of the non treated image # using Otsu algorithm threshold = graythresh(imgNonTreated); # create a binary image based on the grayscale images imgBWTreated = im2bw(imgTreated, threshold); imgBWNonTreated = im2bw(imgNonTreated, threshold); # count foreground pixels for each image and output it in console foregroundPixelCountTreated = sum(reshape(imgBWTreated, 1,numel(imgBWTreated))); foregroundPixelCountNonTreated = sum(reshape(imgBWNonTreated, 1,numel(imgBWNonTreated))); display(strcat("Foreground pixel count for Treated specimen: ",num2str(foregroundPixelCountTreated))) display(strcat("Foreground pixel count for Non-Treated specimen: ",num2str(foregroundPixelCountNonTreated))) ``` ## Functions Guide #### ImageTools Module * _graythresh(image)_ - calculates a threshold of a grayscale image, which can be used downstream to convert a grayscale image to binary image. One input argument is required - image (expected image format obtained by Images.imread, see [Images](https://github.com/timholy/Images.jl) package for more information). E.g. to convert an array to image use grayim(array) or colorim(array). Based on the original Paper: N. Otsu, "A Threshold Selection Method from Gray-Level Histograms" 1979. * _im2bw(image, threshold)_ - converts image into a binary image. Input: image (expected image format obtained by Images.imread, see [Images](https://github.com/timholy/Images.jl) package for more information) and threshold (floating point number). E.g. to convert an array to image use grayim(array) or colorim(array). * _imshow(image)_ - a wrapper for the function _ImageView.view()_. * _imread(path)_ - a substitution function for Images.imread(), which supports reading an image from an URL (similar to Matlab/Octave). * _bwlabel(image,connectivity)_ - labels connected binary regions passed as the _image_ argument with the _connectivity_ defining [pixel connectivity](https://en.wikipedia.org/wiki/Pixel_connectivity) (e.g. 4 or 8). * _jet(numberOfColors::Int64)_ - generates Matlab/Octave-like jet colormap with specified amount of colors. * _hsv(numberOfColors::Int64)_ - generates Matlab/Octave-like hsv colormap with specified amount of colors. * _label2rgb(labeledMatrix, inputColorMap, backgroundColor, isShuffled)_ - creates an RGB image from a matrix of labeled connected regions (produced by the function _bwlabel()_) using following input parameters: matrix of labeled connected regions _labeledMatrix_, _colormap_ to be used _inputColorMap_ (e.g. _inputColorMap = "jet"_), background color in the RGB image produced _backgroundColor_ (e.g. _backgroundColor = [1 1 1]_), _isShuffled_ (e.g. _isShuffled = "noshuffle"_) #### Support Module * _mat2im(array)_ - an auxiliary function that converts a matrix (i.e. Julia array) to an image in the format of _Images_ package * _im2mat(image)_ - an auxiliary function that converts an image in the format of _Images_ package to a matrix (i.e. Julia array) * _rosetta(jlFilePath [, jlFilePath])_ - an auxiliary function for preliminary conversion of m-files to jl-files. _rosetta()_ takes the path to the m-file from the first input argument, replaces the Matlab/Octave comment symbols by Julia comment symbols and injects the _using MatlabCompat_ statement to facilitate functions namespace compatibility. If only one argument is provided the parsed code is returned by the function as a Dict. If second argument is provided _rosetta()_ will save the parsing results into provided path. #### Io (Input/Output) Module * _load(matFile, [treatAs, variable])_ - a Julia analogue of the Matlab/Octave function _load()_. First argument defines the path to the mat file. Second optional argument defines whether the input will be treated as either _"-mat"_ or _"-ascii"_ file (e.g. _treatAs = "-mat"_). Returns Dict containing the mat-file contents. With the third optional input argument set to _variable = "all"_ returns all the variables. Provide specific variable to return only one variable. default values for the second and third argument are _treatAs = "-mat"_ and _variable = "all"_ respectively #### StringTools Module * _num2str(number)_ - covert number to a string in a Matlab/Octave style. * _disp(string)_ - wrapper for Julia _writeln()_ function for Matlab/Octave compatibility. * _strcat(stringsToAdd, ...)_ - wrapper for Julia _string()_ function for Matlab/Octave compatibility. #### MathTools Module * _numel(matrix)_ - wrapper for Julia _length()_ function for Matlab/Octave compatibility. * _max(vector)_ - wrapper for Julia _maximum()_ function for Matlab/Octave compatibility. #### Copyright © 2014-2015 Vardan Andriasyan, Yauhen Yakimovich, Artur Yakimovich