Wavelet-based Image Compression And Edge Detection Research

In this thesis, we not only introduce the basic theory of wavelet but also study the image's characteristics of wavelet coefficients. And we also extend the conception of zerotree by analyzing the zerotree structure of an image. We propose a new conception of "generalized zerotree", and then , We improved the EZW algorithm by the generalized zerotree. The results of experiments indicate that the new algorithm increases the ratio of compression and improves the SNR on the fixed compression ratio.This paper also discusses the optimum criterions of edge_detection, points out that the derivative of Bspline function can approach the optimum dege_detector and defines the B_spline wavelet by the derivative of B_spline function. And then an edge_detection experiment is carried out by the B_spline wavelet transform and it's fast Mallat algorithm.