IMO speech recognition combined with natural language processing will bring the biggest innovation in gaming. Although that might still take a few generations.
Being able to have a conversation with characters in game just as if you are playing online co-op will be more immersive then stereoscopic 3D and full body motion control combined. Just imagine Zelda with an AI controlled navi you can talk to and ask questions to help figure out what to do next. The main character finally doesn't have to be a mute anymore! No more picking from canned responses, although in the beginning most responses will be, can you rephrase that or I can't help you with that.
Another interesting thing to do with the camera is track the emotions of the player. Pick up whether they are frustrated, angry, sad, excited or bored and adjust the game accordingly. An AI director that can read and manipulate your emotional state would be awesome.
