Adding some sort of video animation support. As in, the program motion tracks the video from (for example) frame 20 to frame 30; according to the user's input. Then, the user applies the reference points for the facial animation, thus giving us the most powerful facial animation tool in the world :D

or an alternate solution, use a 3rd part motion tracking program, import the data to CrazyTalk and CrazyTalk applies full 3D facial animation according to the tracking data.

I know I'd love to have this feature since I'm making a machinima film and I need to somehow make the speech look realistic.