You said "the only visual data that needs to be sent across network is the positioning of facial features and hand movements". To extract facial features and ha

You said "the only visual data that needs to be sent across network is the positioning of facial features and hand movements". To extract facial features and hand movements is non-trivial and requires video analysis and may hence involve lag.