Hmm. The SAT seems to heavyweight for what I'm thinking.
I'm trying to find a good 2D graphics clipping algorithm. The view region can be moved and rotated, as can the entities that would appear in it.
The method I currently use is "lossy" - I have my viewing region, and I check if each entity's location is within its bounding radius away from the view. That way I don't have to do any complex math to determine if it should be rendered or not. However, I want to look at other, more specific methods, to see what else is possible.
Here are some examples:
As you can see, object A will be rendered:
However, object B will not: