Render Targets and analyzing the image


I found tutorials on how to paint on CanvasRenderTargets and I got some questions:
I want the player to draw a picture. This picture then should be stored in a (2DTexture ?-) variable.
After that this picture needs to be compared with another picture to check, whether the player has drawn the right (almost same) picture.
Is it somehow possible ?

Would it be also possible to make the player draw in some kind of grid where his inputs are snapped to a custom grid?