"sensitising" an image

Hello all,
I've got a big image that I'd like to use as the UI in an application. A small part of the thing is attached.

What's the best way of senseitising it?
Ideally, I'd like to click on the image somewhere and change that small part of the image (by covering it with another smaller image OR with a small widget - giving the appearance of something changing state).

I assume I'll have to cover all of my "active" regions with widgets or images. which is best?
What's best for tracking the active areas if the user resizes the top widget? Should I just scale the original "measured" map, or should I go looking for things in the original image?

Or should I make a "farm" of widgets instead of using the original image?
What's the best way of doing this? any ideas?

Thanks
K

IMO it's just the perfect place to use the canvas architecture (either QCanvas from Qt3, QtCanvas (available from TT ftp site) from Qt4 or wait for QGraphicsView when Qt 4.2 is released).

As wysota this would probably best be solved by first drawing the image on some widget and then drawing other, smaller, images on top of that image when the user clicks. The advantage of this is that you can easily map out the coordinates of the places where the image shall be changed in a image-editor.