Color Video Compression Based on Chrominance Vector Quantization

This paper proposes a compression technique to improve the quality of color in very low bit rate coding of video. The general idea is to convert the two chrominance components to one scalar chrominance which is processed further. The scalar representation of chrominance is obtained through vector quantization in the chrominance plane. Each (CB ; CR) vector is represented by a scalar index to a chrominance codebook. Experimental results show that coding of this scalar signal improves visual quality of frames both in intraframe and interframe modes as compared to standard codecs.