You can test effectiveness of your diagnose algoritm and optimize it with Matlab Image Processing Toolbox.

After this step, you can write on Java / C++ . As I assume that you will deal with images, you should implement your application with OpenCV by using Android NDK tool. I think, it will make your process speed faster.

Thank you Mr. Ömer Faruk

You can use a Java within Matlab. The link below describes this approach:

http://www.mathworks.com/help/matlab/using-java-libraries-in-matlab.html

Besides use OpenCV and Android NDK tool, you can implement your own algorithms with the use of RenderScript in Android API:

http://developer.android.com/guide/topics/renderscript/compute.html

Just make sure you don't mess up the rows/columns. A row in java is a column is Matlab.

Using C/C++ is fully supported by matlab. You can use the "mex" compiler to build modules directly usable by matlab. You only need to observe certain conventions, such as naming your module in an appropriate way and writing a specific entry point function. See attached link for further instructions.

Cheers,

Fernando

http://www.mathworks.es/es/help/matlab/ref/mex.html

Dear Fernando,

Excellent link. I would also like to work with C linked to MATLAB. Can you please suggest any entry point site for beginers other than MathWork site?

Have you tried Octave ?

Hi Barbara, i don't try it.

 C/C++ is a good to be combined with Matlab. you may as well  use Octave.