Video coding for ISDN applications using predictive classified vector quantization

A video coding algorithm based on predictive classified vector quantization (PCVQ) is described. For teleconferencing applications, it achieves bit per pixel (BPP) values of 0.1856 for the Y plane, 0.0811 for the U plane, and 0.0855 for the V plane of a commonly used color image sequence called Miss America with acceptable subjective quality. This is equivalent to a bit rate of 89.262 kb/s for a frame size of 144*176 pixels and a frame rate of 10 frames/s. Together with its simple structure, these results demonstrate that it can be used for video transmission over ISDN (integrated services digital network) channels.<<ETX>>