Significance tree image sequence coding with DCT-based pyramid structure

Abstract

This paper addresses a new DCT-based embedded significance tree image sequence coding for progressive image transmission. In order to maintain compatibility with the conventional standardized algorithms and achieve an improvement of coding efficiency at very low bit rate, the MC/DCT hybrid image coding structure is fully utilized. Rearranging the DCT transformed coefficients into significance tree structure, and using a modified embedded zerotree quantizer, a new image sequence coder can be built, from which we can achieve an excellent coding performance for progressive image sequence transmission and precise user-specified rate control. In the simulations, the proposed new image coder can give better rate-distortion performance than that of wavelet-based one.