Progressive Image Coding using Augmented Zerotrees of Wavelet Coefficients

Most state-of-the-art image compression techniques exploit the dependencies between the subbands in a wavelet transformed image. We propose a progressive image coding technique which is based upon the augmented zerotrees of wavelet coefficients. An augmented zerotree can be regarded as a spatial orientation tree with some extended characteristics. Both the coder and decoder require to maintain only one list. The augmented zerotrees are constructed in an efficient recursive manner and the algorithms are provided. A comparison reveals that the coder performance is close to, and sometimes outperforms, other well known coding methods. Suggestions are made for further improvements regarding the speed and compression performance of the coder.