High-Performance Embedded Morphological Wavelet Coding

Morphological analysis can be applied in wavelet domain to analyze and represent the position of significant coefficients. New operators have to be introduced which are able to exploit both the multiresolution and the filter bank peculiarities of the subband representation of visual information. In this paper an efficient morphological wavelet coder is proposed. The clustering trend of significant coefficients is captured by a new kind of multi resolution binary dilation operator. The layered and adaptive nature of this subband dilation makes it possible for the coding technique to produce an embedded bit-stream with a modest computational cost and state-of-the-art Rate-Distortion performance. Morphological wavelet coding appears promising because the localized analysis of wavelet coefficient clusters is adequate to capture intrinsic patterns of the source which can have substantial benefits for perceptual or even object-based reconstruction quality concerns. Here we test the performance of our algorithm and compare the effects of different wavelet filters. We obtain state of the art coding performance and good perceptual results both for 2D and 3D images, with a new technique that seems to be well suited for further developments.