Dual level intraframe coding for increased video telecommunication bandwidth
David M. Saxe, Richard A. Foulds, Arthur W. Joyce · 1998
While digital video transmission and video conferencing methods have improved significantly over the last few years, the transmission of sign language for individuals who are deaf via this medium still remains a problem. Desktop video teleconferencing systems accommodate the bandwidth limitations of both analog and digital (ISDN) telephone channels by reducing the frame rate while preserving voice quality and only minimally degrading image quality. Sign language transmission requires fidelity to movement (consistent and high frame rate), and requires reasonable image quality only in the areas around the hands and face. This paper presents a dual-level compression approach which uses a newly developed technique to identify the hands and face from the remainder of each video frame. This allows for a very lossy, high compression of most of each frame, while retaining the visual quality necessary to identify hand shapes and read facial expressions. By taking advantage of this compression, additional bandwidth is recaptured to allow an acceptable frame rate that maintains the fidelity of human movement necessary to represent sign language.