Large block entropy-constrained reflected residual vector quantization

Multispectral imagery and video coding applications benefit from the use of large vector sizes. Other applications also require large vector sizes such as variable dimension Vector Quantizers (VQ) and transform VQ. Entropy-constrained reflected residual vector quantization (EC-RRVQ) is an algorithm that is used to desgin codebooks for image coding with large vector sizes in addtion to high output rate while mainting a very low complexity in terms of computations and memory requirements. EC-RRVQ has several advantages which are important. It can outperform entropy-constrained residual vector quanization (EC-RVQ) in terms of rate-distortion performance, encoder complexity compuatations, and memory. Experimental results indicate that good image reproduction quality can be accomplished at relatively low bit rates. For example, a peak signal-to-noise ratio of 29 dB is obtained for the 512 x 512 image LENA at a bit rate of 0.2 bpp with dimension of 16 x 16.