Given the pixel densities of cameras, I was wondering if it would make sense to capture a very wide field of view, and then extract head and shoulder shots (pos

Given the pixel densities of cameras, I was wondering if it would make sense to capture a very wide field of view, and then extract head and shoulder shots (possibly multiple). The distortion could be untangled with the GPU. Face tracking is getting pretty damn good now. For bonus points call out to web service for recognition and tag each stream appropriately.