Chiralization/mirroring (new functionality)

I'm asking you to add a variant analysis that can measure Aus of face sides independently.
Re-training the whole network would be a hassle, but there's an easier method.
The rough algorithm is as follows: take a detected face, mirror it across a middle line, taking the pose into account, then extract AUs for the original and both reflections separately.

Here's the general idea - the face aligned in Openface and mirrored images on both sides.

It can also be seen that here the alignment is rather poor, but mirroring was done manually, without using facial landmarks.

The issue names no files, tests, or entry points. Start by locating the face-alignment and action-unit extraction paths, then determine how pose-aware mirroring could produce the original and two reflected faces independently. Done means the variant analysis reports separate AUs for all three images without retraining the network.

Written by the indexing model from the issue text.