Predictive Grayscale Image Coding Scheme Using VQ and BTC

A predictive image compression scheme that combines the advantages of vector quantization and moment preserving block truncation coding is introduced in this paper. To exploit the similarities among neighboring image blocks, the block prediction technique is employed in this scheme. If a similar compressed image block can be found in the neighborhood of current processing block, it is taken to encode this block. Otherwise, this image block is encoded either by vector quantization or moment preserving block truncation coding. A bit-rate reduced version of the proposed scheme is also introduced. According to the experimental results, it is shown that the proposed scheme provides better image quality at a low bitrate than these comparative schemes.