Wavelet-based compression methods maintaining multiple lossless region of interest

Images, used in several fields such as medicine or video transmission, contain regions that are more important than others. Compression methods capable of delivering higher reconstruction quality for the important regions are thus attractive. For such cases, only small portions of the image might be essential to the user, and the cost of a wrong interpretation could be high. Methods that deliver lossless compression within the regions of interest, and lossy compression elsewhere in the image, could be a solution, providing an efficient and accurate image coding. In this paper, we propose two methods, which achieve lossless compression of the regions of interest. Both are based on the embedded zerotree wavelet coding which is used according to two different approaches. In one of these approaches, our method is applied to the whole image, while in the other, it is applied to a subdivision in equal blocks of the image.