Using MATLAB shared library in visual studio c++

Question

I have a m-file in MATLAB.It calulates normxcorr2 between an image and a template as follow:

function norm_res = TMatch(temp,image)
    c = normxcorr2(temp,image);
    norm_res = max(c(:));

when I running this m-file from MATLAB command line with two images as image and template, it gives me 0.8076 as a result. Now I build it as a shared library with following command:

mcc -B cpplib:TMatch TMatch.m

in MATLAB command line and it generates required files for using in C++.

Now I use generated shared library in visual studio as following:

    mclInitializaApplication();
    TMatchInitialize();
    Mat img = imread("image.jpg", IMREAD_GRAYSCALE);
    Mat tmp = imread("temp.jpg", IMREAD_GRAYSCALE);
    // these files are the same as files which are used in MATLAB

    mwArray mwImg(img.rows,img.cols,mxDOUBLE_CLASS,mxREAL);
    mwArray mwTemp(tmp.rows,tmp.cols,mxDOUBLE_CLASS,mxREAL);

    mwArray result;

    mwImg.SetData(img.data,img.cols * img.rows);
    mwTemp.SetData(tmp.data,tmp.cols * tmp.rows);

    TMatch(1,result,mwTemp,mwImg);

    double normed_res;
    result.GetData(&normed_res); //finally result will be 0.198949

As you can see when I run m-file in MATLAB result will be:

0.8076

and when I generate a shared library and use it in visual studio C++ the result will be:

0.198949

please tell me why this happens?
Thanks in advance.
(@Miki @A.Riazi)

herohuyongtao · Accepted Answer · 2015-08-31 14:30:08Z

1

The reason is that MATLAB and VC++/OpenCV use different memory management strategies for matrices, where MATLAB is column-major while row-major for VC++/OpenCV.

Try to transpose the matrix in MATLAB or in VC++/OpenCV before calling MATLAB.

answered Aug 31, 2015 at 14:30

herohuyongtao

50.9k30 gold badges138 silver badges179 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

s.a.t Over a year ago

I'm so grateful for your answer @herohuyongtao. I add following lines of codes to my VC++ program: Mat t_img, t_tmp; traspose(img,t_img); traspose(tmp,t_temp); and run the previous code except replace "img" and "tmp" with "t_img" and "t_tmp" and run again but nothing changed and I recieve previous answer again (0.198949). I tried my MATLAB code with following lines: temp = transpose(temp); image = transpose(image); and again I do "normxcorr2", in this case again I recieve previous answer (0.8076 ) and nothing changed. Can you explain to me if I do something wrong?

Collectives™ on Stack Overflow

Using MATLAB shared library in visual studio c++

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related