A Markov system for vector quantization coding

In the present paper, the Markov system for VQ is introduced to exploit the interblock correlation by statistical processing and using different coding orders, like zig-zag, peano, row-major and column-major scanning orders. Computer simulations are carried out to many different natural images. The proposed method is shown to achieve a better bit-rate reduction than Address-VQ with less computational complexity and lower memory requirements. Moreover, the quality of reconstructed image is the same as conventional VQ. Therefore, the proposed strategy raises the compression efficiency greatly.