If you would like to save your money for using a Matlab license (which is a substantial amount) you can go for an open source tool by using Python + image processing tools. You can have a look at this if you are interested.

http://www.linuxforu.com/2012/04/exploring-software-scientific-python-and-image-processing/

It should work fine on all platforms.

Yeah..u cn

I think you mean protein separation images. Sure Matlab can do it. You need determine a threshold to identify 2 D peaks. Then generate a database about the locations of the center of the peaks. Basic idea is to compare the pixel intensities with the threshold value and then group the adjacent pixels to a same patch. Then either use center-of-mass or a 2-D Gaussian fit to find the center. Google "matlab 2d Gaussian fit image" should give some information about the codes.

Sure. You can program any function (images are handled as matrices) but in most cases these functions are already provided. There are several image processing tutorials on the web.

Prediction is something appart. According to my limited knowledge you may, for instance, recognize a shape (in fact by a built-in function), track its position along a series of frames, and estimate its future trajectory.

Interpretation and prediction depends on your knowledge about the problem.

One more critical factor to mention is the computation time and memory.

Consider using CUDA-toobox and/or parallel processing toolbox..

If you would like to save your money for using a Matlab license (which is a substantial amount) you can go for an open source tool by using Python + image processing tools. You can have a look at this if you are interested.

http://www.linuxforu.com/2012/04/exploring-software-scientific-python-and-image-processing/

It should work fine on all platforms.

Yes, you can use it. You can convert the 2D into multiple vectors and use. If not there are plenty of functions anyway.

Clearly you can analyze any image - thus also 2D images. Matlab is truly excellent for image processing. The interpretation will be dependent on your access to library values for peaks. You will have to consider warping when you compare your data with library data. It is not clear what you mean by predict results taken from 2D page.

Of course we can use it as a processing tool ....

I think we can use PCA library tools to reveal and predict information

I agree with Oliver Tomic's answer

Dear Narendrakumar, Are you going to convert 2D image to be 3D image? It is not clear for me. However, I can say MATLAB is very powerful (without any intention to put aside many other tools/language) for digital image processing works.

Is it handwriting recognition or printed text recognition?

Thank you all for your answers... This has put some light on my work. Which i can take it forward..

Sir, in mathlab for image processing as far as i have tried , using the wavelet transformation technique gives brilliantly amazing results . try wavelet transform sir.. I used it for image processing it does any thing related ti image .. i have a few programs that i have written in math lab for ISRO that i can share with you.. :)

For pictures taken crom a robot you can detect the borders and use kalman filters to predict the trajectory

Well matlab is good to deal with the images by its tool box but its is a proprietary software. If you are well concern towards the open source then Octave can give you the similar flexibility to deal with image processing. Octave has similar feature like matlab and even having the same syntax like matlab.

I refer you to go through these links. It may give you enough information.

http://stackoverflow.com/questions/3319355/alternatives-to-matlabs-image-processing-toolkit

http://octave.sourceforge.net/image/index.html

I am convinced that the answer is yes. If you are having a simple single lane gel, you can easily write a code that will give you a profile, that is a significant improvement over a simple visual observation.

When it comes to 2D gel's - things gets more complicated - the gels tends to shrink and deform as they dry - so you need to spike your sample with compounds that you can easily identify in the 2D gel after the run. You can then use those markers to evaluate how to warp the resulting image in order to make it comparable to a reference library that you may have yourself or find in other labs.