Dynamically change camera alpha mask? How?

I just started exploring this idea and really don't know where to start, so I apologize if the question seems a bit vague.

Heres what I would like to do.
Before



After


Basically, where ever the "player" moves, it reveals a picture under it. I know I can set up two camera to render seperate things, but I'm not sure how to create a mask, and then how to make that mask dynamic to reveal things dynamically.

One example I see this used is in some coloring apps for kids. Where ever the screen is colored on, it basically shows and already colored/texture version under it.

Any guidance on getting moving on this would be wonderful.

Thanks!