Video-enabled cell phones have the potential to enable deaf Americans to speak in their community's native language, American Sign Language (ASL), and gain the freedom, flexibility, and comfort of the wireless phone revolution. This research involves the design, implementation, and evaluation of new standards-compliant data compression methods that will allow ASL video, and other structured video, to be transmitted over low bandwidth cell phone channels. To be more specific, the goal of this research is to develop a framework in which to implement low-complexity H.264 encoding to provide maximum quality compressed structured video. Because the developed algorithms will be H.264 standard-compliant, an off-the-shelf H.264 decoder can be used.
The research applies to any class of videos where there is some inherent structure that can be exploited. The five areas of research are: (1) Design of an appropriate objective distortion metric for ASL; (2) Design of algorithms to preprocess video for display on small devices; (3) Development of methods to exploit structure in video for efficient coding; (4) Implementation of a rate-distortion-complexity optimization of H.264 using the new metric; and (5) Applications of the framework to other structured video such as surveillance video. Studies with ASL users are used to develop the objective intelligibility metric and to evaluate the new compression techniques.
