Old Approaches

My 1st ever method was just to use a single linetrace in the dead center of the screen, but that would require precision shots.

My 2nd method was to use a big box trace that would extend quite far in the forward direction. It worked, but the problem was any animal that would be near the player (not even in the frame) would get hit before the one in the in the camera shot.

My 3rd approach was to use a cone mesh and check what ever overlaps with. Quite a simple method but unfortunately it wasn't working properly. It would detect some animals but not the others. Another problem was checking if the animal is on the screen as well and not behind trees, rocks etc

Current System

Its an iterative box trace approach. The size grows bigger each iteration. Its not the best system because this approach uses For Loop, that in the worst case can go up to 150 iterations. It only checks for pawns and doesn't collide with anything else.
then a line trace is done to the hit animal to check whether the player can even see this animal. If so, the picture gets tagged with the animal's name and saved onto both the disk (as a png) and into the save game (the path at which this png can be retrieved).