Wavelet image coding by dilation-run algorithm

Abstract

This paper presents a novel wavelet image coder, dilation-run algorithm, which provides an embedded coder based on bit-plane coding technique. At each bit-plane, morphological dilation is used to extract and encode the clustering significant wavelet coefficients, and a run-length coding method is used to encode the positional information of each cluster. The experiment results show that our coder outperforms the zerotree coder SPIHT and is competitive with the morphology coders MRWD and SLCCA. For wavelet images with strong clustering feature, the new coder outperforms both the morphology coders above.