A Unified Scheme for Fast EZW Coding Algorithm

Embedded Image Coding Using Zerotrees of Wavelet Coefficients (EZW) is a common and efficient image encoding algorithm. Many authors have improved it, but they aren¿t able to completely reduce redundant search, they have not fully improved the coding speed. So Shapiro advanced zerotree map technique to eliminate redundant search in 1996. A typical example using zerotree map is Mow-Song Ng¿s EZW coding system. But in Mow-Song Ng¿s algorithm, the coefficient-nodes in dominant list and subordinate list have to contain more information, reducing the coding speed. Thus, this paper propounds a unified scheme for fast EZW encoding algorithm and decoding algorithm, it encodes or decodes on an ordered linear list and its coefficient-nodes contain less information, improving the encoding or decoding speed to a certain degree. Many EZW encoding algorithms using different scan order, including Mow-Song Ng¿s algorithm, can be transformed into this unified scheme.