The Application Of Zero-Trees In Image Compression

The wavelet analysis and the wavelet transform are new tools of signal processing developed in later period of 80's. Because of their good local performances in both time (space) and frequency, they are particular for the analysis of non-stationary signals and have been used in various fields of signal processing recently. In this dissertation, the still image compression based on wavelet transform is studied.Although wavelet transform is efficient for image processing, different wavelet bases have different efficiency for image coding. Our research shows that linear phase, prefect reconstruction and regularity biorthogonal wavelet filter banks are better than orthonormal wavelet filter banks for image compression.Zero-trees coding is one of the most efficient still image compression , this paper focused on the discussion about zero-trees in the lossless image compression, and analysis the three parameters—the choose of wavelet based, the rest matrix coding and the layer of zero-trees—which effected the efficiency of zero-trees coding and gave the means of how to choose the parameters. Wavelet decompose can made the energy of image get together to the low frequency, thus it can form more zero-trees and MSR as the measure of the efficiency of Zero-trees coding. The layer of the Zero-trees coding can be got by the minimum bits of Zero-trees coding. The rest matrixes R coding adopts the math coding though statistics methods.