Abstract
The image compression algorithms using wavelet transform has demonstrated their elegant performance. The wavelet packets present a generalization of the multiresolution analysis and could perform adaptive transform on the whole image. In this thesis we proposed an image coding algorithm using wavelet packet. Besides, the compressed bit stream has the embedded property, which is naturally useful for progressive transmission. The bottom-up best basis selection algorithm based on entropy criterion make our wavelet packet representation always compact. The combination of data reordering and successive approximation quantization with arithmetic coding promise to generate an embedded bit stream. Experiment result shows that our proposed method outperform than well-known wavelet-based coder, such as EZW. For image containing larger high frequency component, our wavelet packet coder could preserve more characteristic of the image than those using wavelet transform.