http://stackoverflow.com/questions/1607933/running-matlab-function-from-java

Running MATLAB function from Java

up vote11down votefavorite 7

I have an .m file in MATLAB which I would like to call from Java an get the solution as a string or whatever in Java. This sounds really simple but for some reason I can't make it work. I tried this: matlab -nosplash -wait -nodesktop -r myFunction but I'm not sure how I parse the answer since MATLAB opens it's own command line (in Windows). I use this, but it doesn't return anything. Process p = Runtime.getRuntime().exec(commandToRun); BufferedReader stdInput = new BufferedReader(new InputStreamReader(p.getInputStream())); also it seems that every time I call MATLAB it opens a separate window which is a problem because I'd like to run this many times. java matlab share|improve this question edited Oct 23 '09 at 3:11 Amro 105k19171311 asked Oct 22 '09 at 15:13 Tom 56114

perhaps -logfile command option might help: stackoverflow.com/questions/1518072/… – Amro Oct 22 '09 at 16:28 add a comment |

5 Answers 5

active oldest votes

up vote7down vote	The trick is to use the MatlabControl class http://www.cs.virginia.edu/~whitehouse/matlab/JavaMatlab.html. It's very easy to use and you can do exactly what you're trying to do (and more). share|improve this answer answered Oct 23 '09 at 2:53 Jeff Storey 30k40160314
	Jeff! Rob! Good to see you guys on here. – Scottie T Oct 23 '09 at 17:16 Small world, isn't it? – Jeff Storey Oct 23 '09 at 17:20 add a comment |

up vote7down vote	matlabcontrol is based on the same underlying MATLAB library used by MatlabControl mentioned by Jeff, but is more up to date, reliable, and documented. To get started, take a look at the walkthrough. share|improve this answer answered Jul 8 '11 at 23:20 Joshua Kaplan 64245
	add a comment |

up vote2down vote	JAMAL is an open source, Java RMI-based (Java Remote Method Invocation API) library that suits your needs share|improve this answer answered Aug 19 '13 at 14:41 Maksim Khadkevich 692
	This link do not work. – user1850484 Sep 13 '16 at 10:34 add a comment |

up vote2down vote	In Matlab R2016b, MathWorks added MATLAB Engine API for Java which allows to execute MATLAB code from Java. share|improve this answer answered Sep 15 '16 at 15:32 Quepas 6201622
	add a comment |

up vote0down vote

There exists a good Java-COM-Bridge called JaCoB (http://sourceforge.net/projects/jacob-project/) which you can use to automatically start Matlab as a COM-Server in the background. You can then follow the instructions in the Matlab help to interact with the Matlab COM Interface. Although this is a very generic interface, it provides enough flexibility to easily do a few calls to Matlab like in your case. Simply download the JaCoB package and look in the docs folder for some documentation. You also have to include the Jacob DLL in your path. share|improve this answer answered Oct 23 '09 at 17:14 Wolfgang Ulmer 1,006910

add a comment |

Your Answer

dist_func.m

matlab dist function

de2bi1.m

matlab de2bi source code

The dist function is a 'Euclidean distance weight function' which applies weights to an input to get weighted inputs. At your example:

W is the (random) weight matrix. P is the input vector Z is the weighted input

If you type in the matlab prompt 'edit dist.apply' you find the formula behind this function. For your example, the weighted matrix is subtracted from the transposed and copied vector. Now it is squared and then the square root is taken. This is how the Euclidian norm is defined Norm = square((a-b)^2)

I have copied the code and made the example simpler to understand in the code below better:

clc;clear all;close all; 

p=[1;2;3]
w=[1 1 2;1 1 1;1 2 1;1 1 1]
z1=dist(w,p)

% dist function
S = size(w,1);
Q = size(p,2);
z2 = zeros(S,Q);
if (Q<S)
  p = p';
  copies = zeros(1,S);
  for q=1:Q
    z2(:,q) = sum((w-p(q+copies,:)).^2,2);
  end
else
  w = w';
  copies = zeros(1,Q);
  for i=1:S
    z2(i,:) = sum((w(:,i+copies)-p).^2,1);
  end
end
z2 = sqrt(z2)

Here z1 and z2 should give the same answer.

I hope this makes it clearer.

Kind regards, Christiaan

신호처리 알고리즘을 개발 함에 있어, 유명한 툴 중 하나가 matlab이라는 것이 있다.

지금은 오히려 matlab이 없으면 신호처리 알고리즘 개발 외,

여러 분야에서 개발이 불가능한 상태라고 해도 과언이 아니다.

그러나, matlab 은 고가의 툴로 학생이나 일반적으로 사용하기 곤란하다.

거기에다. 툴박스별로 라이센스비용이 있고 그렇다.

이러한 matlab를 대용할 수 있는 무료의 툴들이 몇가지 존재한다.

그러한 툴중, 본인이 몇가지 사용한 결과 결국은 octave로 귀착되었다.

matlab과 같은 기능과 연산속도를 기대하지 않고 사용한다면,  

matlab과의 문법적 호환성도 좋아 나쁘지 않다. 

최근에는 설치 방법도 이전 버전 대비, installer 한방으로 해결되니,

더욱 접근이 손 쉬워진 것 같다.

[ octave 설치 ]

1. 다운로드
    https://www.gnu.org/software/octave/
    
    원도우용의 인스톨용 파일을 다운로드 한다.
    현시점 (160113) 아래버전이 최신 버전이다.

    
    octave-4.0.0_0-installer.exe    2015-05-28 14:43     175M     
 

2. 설치

   설치파일로 설치 시작하면, 과정중에 BLAS를 선택하게 되어 있는데,

   그냥 default의 openBLAS 

   로 한다.

3. 끝
    터미널 창형식 구동을 위한 CLI 및 matlab 과 유사한 IDE 환경의 GUI 바로가기가 생기는데,

    아직 GUI 버전이 완벽하지 않은 느낌이다. 그냥 터미널 형식으로 사용하면 된다.

    편집기로는 notepad++ 를 권한다 :)

% EOF % 

http://blogs.mathworks.com/steve/2011/09/30/binary-image-convex-hull/

I've been intending to mention a new function bwconvhull that was introduced in the Image Processing Toolbox last spring in the R2011a release. Now that R2011b is out, I figure I better go ahead and do it!

bwconvhull computes the "convex hull of a binary image." Now I have to admit that this terminology is a little loose, so I'd better clarify. The convex hull of a set of 2-D points is the smallest convex polygon that contains the entire set. Here's an example from the MATLAB documentation for convhull:

xx = -1:.05:1; yy = abs(sqrt(xx));
[x,y] = pol2cart(xx,yy);
k = convhull(x,y);
plot(x(k),y(k),'r-',x,y,'b+')

The polygon in red is the convex hull of the set of points shown in blue. So convhull takes a set of points and returns a polygon, whereas bwconvhull takes a binary image and returns another binary image. What's up with that? It means simply that bwconvhull computes the convex hull of all the foreground pixels in the input image, and then it produces an output binary image with all the pixels inside the convex hull set to white. It's a little easier to show than to say, so here's what it looks like:

bw = imread('text.png');
imshow(bw)

bw2 = bwconvhull(bw);
imshow(bw2);

Let's overlay the foreground pixel locations from the input image (using blue dots) and the convex hull computed by convhull (using a thick red line).

[y, x] = find(bw);
k = convhull(x, y);
hold on
plot(x, y, 'b.')
plot(x(k), y(k), 'r', 'LineWidth', 4)
hold off

An important variation supported by bwconvhull is to compute the convex hulls of the individual objects in the input image. Here's how you do that:

bw3 = bwconvhull(bw, 'objects');
imshow(bw3)

Something called the convex deficiency is sometimes used in shape recognition applications. Loosely speaking, the convex deficiency of a shape is the convex hull of the shape minus the shape. Here's an example using the letter "T" from the text image above.

bwt = bw(7:24, 4:18);
imshow(bwt, 'InitialMagnification', 'fit')

bwtc = bwconvhull(bw_t);
imshow(bwtc, 'InitialMagnification', 'fit')
title('Convex hull image')

bwtcd = bwtc & ~bwt;
imshow(bwtcd, 'InitialMagnification', 'fit')
title('Convex deficiency image')

The convex deficiency of the letter "T" has two connected components. This kind of measurement can be useful for recognizing shapes.

https://kr.mathworks.com/matlabcentral/answers/116402-how-do-i-change-the-matlab-user-interface-from-korean-to-english-in-r2014a-and-later

For R2014a and R2014b, see http://www.mathworks.co.jp/jp/help/ko_KR/R2014a/LocaleSwitch.html for details on how to switch the language of the MATLAB User Interface.

Releases from R2015a, you can change language setting in Preference->General->Desktop language from Korean to English

R2014a와 R2014b는 http://www.mathworks.co.jp/jp/help/ko_KR/R2014a/LocaleSwitch.html를 참조하여 MATLAB UI의 언어를 변경할 수 있습니다.

R2015a 버전부터는 기본 설정->일반->데스크탑 언어 설정을 영어로 변경 하시면 영문 UI를 사용하실 수 있습니다.

External Interfaces

Passing Two or More Inputs or Outputs

The plhs[] and prhs[] parameters are vectors that contain pointers to each left-hand side (output) variable and each right-hand side (input) variable, respectively. Accordingly, plhs[0]contains a pointer to the first left-hand side argument, plhs[1] contains a pointer to the second left-hand side argument, and so on. Likewise, prhs[0] contains a pointer to the first right-hand side argument, prhs[1] points to the second, and so on.

This example, xtimesy, multiplies an input scalar by an input scalar or matrix and outputs a matrix. For example, using xtimesy with two scalars gives

• x = 7;
  y = 7;
  z = xtimesy(x,y)
  
  z = 
       49
  

Using xtimesy with a scalar and a matrix gives

• x = 9;
  y = ones(3);
  z = xtimesy(x,y)
  
  z = 
       9      9      9
       9      9      9
       9      9      9
  

This is the corresponding MEX-file C code.

• /*
   * =============================================================
   * xtimesy.c - example found in API guide
   *
   * Multiplies an input scalar times an input matrix and outputs a
   * matrix. 
   *
   * This is a MEX-file for MATLAB.
   * Copyright (c) 1984-2000 The MathWorks, Inc.
   * =============================================================
   */
  
  /* $Revision: 1.10 $ */
  
  #include "mex.h"
  
  void xtimesy(double x, double *y, double *z, int m, int n)
  {
    int i,j,count = 0;
    
    for (i = 0; i < n; i++) {
      for (j = 0; j < m; j++) {
        *(z+count) = x * *(y+count);
        count++;
      }
    }
  }
  
  
  /* The gateway routine */
  void mexFunction(int nlhs, mxArray *plhs[],
                   int nrhs, const mxArray *prhs[])
  {
    double *y, *z;
    double x;
    int status,mrows,ncols;
    
    /*  Check for proper number of arguments. */
    /* NOTE: You do not need an else statement when using
       mexErrMsgTxt within an if statement. It will never
       get to the else statement if mexErrMsgTxt is executed.
       (mexErrMsgTxt breaks you out of the MEX-file.) 
    */
    if (nrhs != 2) 
      mexErrMsgTxt("Two inputs required.");
    if (nlhs != 1) 
      mexErrMsgTxt("One output required.");
    
    /* Check to make sure the first input argument is a scalar. */
    if (!mxIsDouble(prhs[0]) || mxIsComplex(prhs[0]) ||
        mxGetN(prhs[0])*mxGetM(prhs[0]) != 1) {
      mexErrMsgTxt("Input x must be a scalar.");
    }
    
    /* Get the scalar input x. */
    x = mxGetScalar(prhs[0]);
    
    /* Create a pointer to the input matrix y. */
    y = mxGetPr(prhs[1]);
    
    /* Get the dimensions of the matrix input y. */
    mrows = mxGetM(prhs[1]);
    ncols = mxGetN(prhs[1]);
    
    /* Set the output pointer to the output matrix. */
    plhs[0] = mxCreateDoubleMatrix(mrows,ncols, mxREAL);
    
    /* Create a C pointer to a copy of the output matrix. */
    z = mxGetPr(plhs[0]);
    
    /* Call the C subroutine. */
    xtimesy(x,y,z,mrows,ncols);
  }
  

As this example shows, creating MEX-file gateways that handle multiple inputs and outputs is straightforward. All you need to do is keep track of which indices of the vectors prhs andplhs correspond to the input and output arguments of your function. In the example above, the input variable x corresponds to prhs[0] and the input variable y to prhs[1].

Note that mxGetScalar returns the value of x rather than a pointer to x. This is just an alternative way of handling scalars. You could treat x as a 1-by-1 matrix and use mxGetPr to return a pointer to x.

Passing Strings

Passing Structures and Cell Arrays

© 1994-2005 The MathWorks, Inc.

http://www.ninadthakoor.com/2012/03/03/matcv/

opencv mat를 matlab mex 데이터로 변환할때는 .t()로 transpose시켜줘야 한다.

Using OpenCV functions in Matlab

http://stackoverflow.com/questions/27514658/way-to-send-opencv-mat-to-matlab-workspace-without-copying-the-data

When I write MEX files which use OpenCV functions it's easy to pass the data from MATLAB to the MEX environment without copying the data. Is there a way to return the data to MATLAB in the same manner? (That is, without copying the data and without causing MATLAB to crash...)

A simple example:

#include "mex.h"
#include "/opencv2/core.hpp"
void mexFunction(int nlhs, mxArray *plhs[], int nrhs,const mxArray *prhs[])
{

    Rows=mxGetM(prhs[0]);
    Cols=mxGetN(prhs[0]);
    Mat InMat(Cols,Rows,CV_64FC1,mxGetPr(prhs[0]));//Matlab passes data column-wise 
                                                   // no need to copy data - SUPER!

    InMat=InMat.t();//transpose so the matrix is identical to MATLAB's one

    //Make some openCV operations on InMat to get OutMat...


    //Way of preventing the following code??

    plhs[0]=mxCreateDoubleMatrix(OutMat.rows,OutMat.cols,mxREAL);
    double *pOut=mxGetPr(plhs[0]);

    for (int i(0);i<OutMat.rows;i++)
      for (int j(0);j<OutMat.cols;j++)
         pOut[i+j*OutMat.rows]=OutMat.at<double>(i,j);

}

-----------------------------------------

Usually I do the input and output just like that, attaching a pointer to deal with the input and looping over elements on output. But, I think the output can be done in a similar manner to the input, although not without a copy of some sort. The only way to avoid a copy is to create the output Mat with a pointer from a mxArray and operate on it inplace. That's not always possible, of course. But you can be graceful about how you copy the data out.

You can exploit the same trick of attaching a buffer to a cv::Mat that you use (me too!) to bring data in from MATLAB, but also to get it out. The twist on the trick to exporting the data is to use copyTojust right so that it will use the existing buffer, the one from the mxArray in plhs[i].

Starting with an input like this:

double *img = mxGetPr(prhs[0]);
cv::Mat src = cv::Mat(ncols, nrows, CV_64FC1, img).t(); // nrows <-> ncols, transpose

You perform some operation, like resizing:

cv::Mat dst;
cv::resize(src, dst, cv::Size(0, 0), 0.5, 0.5, cv::INTER_CUBIC);

To get dst into MATLAB: first transpose the output (for sake of reordering the data into col-major order) then create an output cv::Mat with the pointer from the plhs[0] mxArray, and finally call copyTo to fill out the wrapper Mat with the transposed data:

dst = dst.t(); // first!
cv::Mat outMatWrap(dst.rows, dst.cols, dst.type(), pOut); // dst.type() or CV_*
dst.copyTo(outMatWrap); // no realloc if dims and type match

It is very important to get the dimensions and data type exactly the same for the following call to copyTo to keep from reallocating outMatWrap.

Note that when outMatWrap is destroyed, the data buffer will not be deallocated because the reference count is 0 (Mat::release() does not deallocate .data).

Possible template (by no means bullet-proof!)

template <typename T>
void cvToMATLAB(cv::Mat mat, T *p)
{
    CV_Assert(mat.elemSize1() == sizeof(T));
    mat = mat.t();
    cv::Mat outMatWrap(mat.rows, mat.cols, mat.type(), p);
    mat.copyTo(outMatWrap);
}

This should be good for channels>1, as long as the size of the MATLAB array is in pixel order too (e.g. 3xMxN). Then use permute as needed.

Note about copyTo

The conditions under which copyTo will reallocate the destination buffer are if the dimensions or data type do not match:

opencv2\core\mat.hpp line 347 (version 2.4.10), with my comments:

inline void Mat::create(int _rows, int _cols, int _type)
{
    _type &= TYPE_MASK;
    if( dims <= 2 && rows == _rows && cols == _cols && type() == _type && data )
        return;    // HIT THIS TO USE EXISTING BUFFER!
    int sz[] = {_rows, _cols};
    create(2, sz, _type); // realloc!
}

So, just make sure you get the size and data type correct, and the data will end up in the mxArraybuffer instead of somewhere else. If you do it right, copyTo will use the buffer you specified, calling memcpy on each row.

아래 설명에 나온것처럼 cmake로 mex파일을 생성할때는, visual studio sln파일을 만든 후, 그 프로젝트를 빌드한 후, INSTALL을 빌드해주면 mex파일이 생성된다.

---------------------------------------------------

Introduction

You will find the complete source code on Stereolabs’ Github page.

If you simply want to use the ZED stereo camera with Matlab (without the ZED SDK) you can refer to this article.

Prerequisites

You must have Matlab installed on your system and of course, the ZED SDK needs to be installed too.

Code

Build procedure

On Linux

Open a terminal in zed-Matlab directory and execute the following command:

On Windows

You need to set the environment variable ‘MATLAB_ROOT’, to do this you can use an editor such as ‘Rapid Environment Editor’ or use the following procedure:

• Click the Start button, right-click on ‘Computer’ and select ‘Properties’.
• Click the ‘Advanced System Settings’ link in the left column.
• In the System Properties window, click on the ‘Advanced’ tab, then click the ‘Environment Variables…’ button in the bottom of the window.
• In the upper part of the window, add your definition to ‘User variables for..’
• Click ‘New…’
• Set ‘Variable name’ as MATLAB_ROOT
• Set ‘Variable value’ to your own matlab root directory e.g. ‘C:\Program Files\MATLAB\R2014b’
• Restart your computer.

Then a common build procedure is used:

• Open cmake-gui.
• In “Where is the source code”, enter the path of the sources (e.g. ‘C:\Users\Peter\Documents\zed-matlab\src’)
• In “Where to build the binaries”, enter the build directory (e.g. ‘C:\Users\Peter\Documents\zed-matlab\build’), and click the “Configure” button.
• A dialog window asks you if CMAKE can create the folder “build” itself. Say yes.
• Another dialog window will ask you the generator of your project. Choose Visual Studio. For example, we choose “Visual Studio 12 2013”. Click the “Finish” button.
• CMAKE may take few seconds to configure the project. Then, some red lines should be displayed in the cmake-gui window.
• Make sure that the message ‘MATLAB Found, MATLAB MEX will be compiled.’ is print, if not the variable environment MATLAB_ROOT can not be found, and you can not compile the project.
• Click the “Generate” button.
• CMAKE has just generated your project in the build folder. Now, you can close the cmake-gui window and go to the build folder.
• Visual Studio files has been generated and a Visual Studio project file named “MEX.sln” too. Open it with Visual Studio.
• Right click on the project named ‘mexZED’ and select ‘Build’
• Right click on the project named ‘INSTALL’ and select ‘Build’

Run the sample

You can now find the generated MEXA file in the matlab directory (e.g. ‘\zed-matlab\matlab’), you can launch Matlab and run the ZED_Camera.m file.

Read More

Error using mex
   Creating library example_mex_function.lib and object example_mex_function.exp
example_mex_function.obj : error LNK2019: unresolved external symbol utIsInterruptPending referenced in function "void __cdecl dlib::check_for_matlab_ctrl_c(void)" (?check_for_matlab_ctrl_c@dlib@@YAXXZ)
example_mex_function.mexw64 : fatal error LNK1120: 1 unresolved externals

이런 에러가 떠서 검색하여 아래와 같이 하니 됐다.

mex -O "example_mex_function.cpp" -I"D:\Research\dlib\dlib-19.0" "C:\Program Files\MATLAB\R2016a\extern\lib\win64\microsoft\libut.lib"

Energy minimization 부분을 살펴보다 관련 소스를 얻어 컴파일하려고 하니 아래와 같은 오류가 발생하였다.

 error LNK2019: unresolved external symbol utIsInterruptPending referenced in function...

utIsInterruptPending가 extern으로 선언되어 있으나 링크되지 않아서 생기는 문제였다.

utIsInterruptPending가 어디에 쓰는 함수인고 하니, Ctrl+C가 입력되었을 때, 이를 처리하는 함수라고 한다.

utIsInterruptPending는 undocumented MATLAB API로 libut.dll에 구현되어 있으며, libut.lib를 mex 컴파일시에 링크해주면 되겠다. libut.lib는 기본적으로 mex에서 링크하지 않고 있으므로 -l옵션을 이용하여 추가해 주어야한다. 다음과 같이 하면 되겠다.

mex blahblah.cpp -lut

출처: http://www.caam.rice.edu/~wy1/links/mex_ctrl_c_trick/

아래의 코드에서

원래는 cc[i*n + j] = aa[i*n + j] + bb[i*n + j];가 comment돼 있으나, 이걸 풀고
//C[i*n + j] = A[i*n + j] + B[i*n + j]; 이것을 comment시켜야 mex가 된다.

To access the variables, you need to associate a pointer to the data in the mxArray. Once you do this, accessing the variables is very simple.


a = mxGetPr(a_in_m);
b = mxGetPr(b_in_m);
c = mxGetPr(c_out_m);
d = mxGetPr(d_out_m);


Now it is possible to access the arrays with standard C or C++ [] notation. There are three important things to remember though:

• You use 0-based indexing as always in C
• You still use column-first indexing like in Matlab, though
• To access the arrays, you use linear indexing (you can’t use [x][y], you have to use [y+x*dimy]

With those three things in mind, go crazy. You can use standard C libraries (as long as you include them). You can use for loops as much as your heart desires, and your code will be much, much faster than its Matlab equivalent.

Note that if we access using C[x] , we will get this error. In other words, we need to define another pointer to access the input memory space using: double *cc = mxGetPr(C);

>> mex mextest.cpp
mextest.cpp
mextest.cpp(36) : error C2036: ‘mxArray *’ : unknown size
mextest.cpp(36) : error C2036: ‘const mxArray *’ : unknown size
mextest.cpp(36) : error C2036: ‘const mxArray *’ : unknown size
mextest.cpp(36) : error C2676: binary ‘+’ : ‘const mxArray’ does not define this operator or a conversion to a type acceptable to the predefined operator

C:\PROGRA~2\MATLAB\R2010B\BIN\MEX.PL: Error: Compile of ‘mextest.cpp’ failed.

??? Error using ==> mex at 208
Unable to complete successfully.

My working code:

#include “mex.h”

void mexFunction(int nlhs, mxArray *plhs[],
int nrhs, const mxArray *prhs[])
{
/* more C/C++ code … */
/* Check for proper number of arguments */
if (nrhs != 2) {
mexErrMsgTxt(“Two input arguments required.”);
} else if (nlhs > 1) {
mexErrMsgTxt(“Too many output arguments.”);
}

const mxArray *A = prhs[0];
const mxArray *B = prhs[1];

/* Check the dimensions of Y.  Y can be 4 X 1 or 1 X 4. */
int m = mxGetM(A);
int n = mxGetN(A);

if (!mxIsDouble(A) || mxIsComplex(A) ) {
mexErrMsgTxt(“YPRIME requires that Y be a 4 x 1 vector.”);
}

/* Create a matrix for the return argument */
mxArray *C;
C = plhs[0] = mxCreateDoubleMatrix(m, n, mxREAL);

double *cc = mxGetPr(C);
double *aa = mxGetPr(A);
double *bb = mxGetPr(B);

for(int i=0; i<m; i++){
for(int j=0; j<n; j++){
//cc[i*n + j] = aa[i*n + j] + bb[i*n + j];
C[i*n + j] = A[i*n + j] + B[i*n + j];
}
}

}

mexFunction내에서 opencv가 라이브러리로 있을 때, mex 어떻게 시키는지 참고할수 있었음.

Normally in MATLAB I can compile a mex file which uses OpenCV functions using:

mex -O "mxFunc.cpp" -I"C:\OpenCV2.1\include/opencv" -L"C:\OpenCV2.1\lib" -lcv210 -lcvaux210 -lcxcore210 -lhighgui210

However, having switched to a 64-bit version of MATLAB, I now get unresolved symbols, e.g.

mxFunc.obj : error LNK2019: unresolved external symbol cvReleaseImage referenced in function mexFunction

How can I fix this?

System: Windows 7 64-bit; MSVC 2005; MATLAB R2010b 64-bit; OpenCV 2.1.0.

-----------------------------------------------------------------------------------

Generally: You need to recompile the used libraries in 64-bit.

As far as I know, it's not enough. If you use STL (and the OpenCV uses a lot) you need to use the same CRT version what the Matlab uses. So you need to use the same version of MSVC what the Mathworks guys...

You can check the dependency of the libmex.dll to figure out which CRT is needed. After it you need to install the proper Visual C++ (normally the free version is enough).

-----------------------------------