Motion-compensated video image compression using luminance and chrominance components for motion estimation

Summary form only given. A variable rate video image compression system using motion-compensated discrete cosine transform (MCDCT) coding is investigated. This paper focuses on the quality improvement achieved by using the luminance component (Y) and both chrominance components (Cr and Cb) for motion estimation. The design of the MCDCT system is based solely on existing technology. The complete system is currently simulated in software. The simulations of the discrete cosine transform processor and the motion estimation processor are based on the technical specifications of existing SGS-Thomson components. The performance of the MCDCT image compression system is evaluated using video images collected from a shuttle mission. The significant error count is shown to be sensitive to the luminance or color distortions in localized areas in an image sequence that are noticeable to viewers, but do not cause significant variations in the SNR measurement.<
>