Finding the actual mouse position is harder than you might think - the various different browsers do it in different ways (and it obviously has to be done at the client). This may help you:
http://docs.jquery.com/Tutorials:Mouse_Position[
^]
But personally, I wouldn't. Why not? Because what are your users going to do when they don't have a mouse? A Tablet, or smartphone for example.
The other approach might be to divide the image into parts: if you make it into nine separate images then it is easier to work with as they just need to touch the appropriate part image - which can have it's own event. Depends how accurate you want your users to be how many images you need.