Object Recognition require image processing techniques, you can design a framework for this which demonstrate your processing steps or else I recommend you to go through the following link.

http://www.cse.usf.edu/~r1k/MachineVisionBook/MachineVision.files/MachineVision_Chapter15.pdf

Dear Chaitali Laulkar,

You mentioned that your problem statement is "Object recognition in an image".

This is not a complete problem statement. You asked how to prove the problem statement. It seems you are asking how to describe the problem statement and how to justify that the problem taken is worthy and reasonable. 

If this is your intention, then you find the following.

Because object recognition in an image is more general, the problem description should make the type of images considered and the type of object to be recognized.

Then formulate the system model which include the initial noise if any and corresponding pre-processing filters. 

Then make an exhaustive study of the algorithms currently in use to recognize the object that we planned in an image we considered. 

Then understand the practical use of those and limitations of the early techniques. 

May be with this you can have a complete problem description.