Abstract
Image compression has been studied for years in order to reduce the representation of the bit streams of an image into a smaller file with less optic loss. Wavelet compression is found to possess more advantages than others while a variety of previous approaches have been proposed in this field. The first part of the thesis consists of a comparison of four schemes of image compression, JPEG, vector quantization, fractal compression, and wavelet compression, all of which are most popular in the past decade. Deeper understanding of them would be built through the survey of some related work and its experimental results. In the other part, this thesis studies a method of the wavelet compression and proposes a modification of SPIHT with RoI. The experiments depict not only the importance of a compression method that provides the primitive RoI functionality at a very high compression ratio over the Internet but also the better performance of this modified scheme in the harsh condition of the Internet.