I've been looking through Graphics, Graphics2D and can't seem to find
what I need.
I'm looking for a method that takes an image and 4 coordinates that
identify a subset of that image.
But all the methods that I find are restricted to a rectangular shape.

Is there a method that takes in an arbitrary 4-sided shape subset of
the original image and outputs those pixels in a rectangular shape?

i.e. I have a 100x100 image.  I want the subset of (0,0),(0,50),
(100,0),(100,25) of the original image to be sampled and outputted to
a new 100x100 image.

Thanks.