Research Of Image Compression Based On Wavelet Transform

Image compression is a rising research field. The break-through in this field will have a profound and far-reaching influence on development of communication and multimedia. Wavelet analysis is a novel field with rapid development in modern mathematics; moreover in has double meanings of deep theory and wide usage. Wavelet image compression is one of the important applications for wavelet analysis. It has advantages of high compression rate, high compression speed, unchangeable signal and image characteristics after compression, and anti-interference under transmitting.In order to develop wavelet transform applied to image processing, this paper presents a theoretical formula to evaluate energy distributions of sub-images in every layer, after wavelet transform, gives a method of defining wavelet coefficient threshold value, and completes the preliminary image compression. Then the further image compression is completed by quantizing the remnant of wavelet coefficient based on clustering analysis. For example, the problem of compressing Lena's picture with 256 X 256 pixels is discussed in detail with the new method of image compression.The emulation result through Matlab software indicates: the computation is simple and the compression rate is high using our method. Comparing with the method of wavelet zero-tree compression encoding, our method is better and more fit in with image compression under high compression rate.