Add depth perception and/or two camera support

[Adam-McDevitt]

Previous iterations of the app support two cameras at once. This mode would assign one camera to each eye, so as to simulate depth perception.

It did... work, but it wasn't particularly good. Heuristics were implemented to prevent the eyes going inappropriately cross-eyed, and to try and make sure they were following the same target, but the eyes moved somewhat independently and often shot off on their own.

It would be good if they eyes supported depth perception again, but better. If I stand close to the screen and between the eyes then it makes sense for them both to look at me (slightly cross-eyed) but in this current iteration of the app they just stare straight ahead.

This could be implemented with two camera support again, but with more effort to prevent the eyes going haywire. Alternatively, depth could be estimated from one camera through some algorithm.

See commit 5c2a93a as the last commit that supported two cameras.