Research On Wavelet-based Image Compression Method

The deficency that traditional fourier analysis tool lacks the ability to locally analyze signal in the time domain and spectral domain simultaneously limits its application in the non-stationary signal realm significantly.However,wavelet analysis has a inherent self-adaptive analysis ability that it has a good resolution in the frequency domain for low frequency signals while providing good resolutions for high frequency signals in the time domain.This attribute provides more flexibilities for signal processing operations.Besides,wavelet can concentrate the energy of signal on several relative large wavelet coefficients,and the wavelets coefficients corresponding to the same location at different decomposition levels are intimately related.These advantages make wavelet the first choice for a variety of applications in signal processing realm,such as signal de-nosing,edge detection,image compression,etc .With the advance of technology and the development of society,many more image applications emerge and the demand for image quality are increasing.Due to the limited bandwidth and storage capacity,image usually has to be compressed before serving various demands.Wavelet has splendid performance with image compression applications,largely due to its unique chracteristics. This dissertation reviews the history of wavelet analysis at the first, followed by an explaination of its characteristics and multiresolution analysis theory in depth.After a review of the background of imaged compression, JPEG and JPEG2000 image compression standard are covered and the performances of these two standards are also compared with a notice of pros and cons of each other. Understanding of the principle of lifting algorithm, we review several proposed architecture for the implematation of CDF9/7 wavelet adopted in JPEG2000 image compression standard. Combining the predict and update procedures, the dissertation proposes a algorithm to implement the CDF9/7 wavelet with the corresponding architecure.Comparisions with other architecture are also given.Finally, looking to the forward, we summary the work and analyze the defects with an expectation of improvements and breakthroughs.