Medical Image Compression Based on Region of Interest and Data Elimination

Compression has been an important part in image processing due to high amount of information contained within an image, reducing it has become a necessity specially in the medical field, where everyday a lot of studies have to be stored and transmitted for medical uses. Nowadays exists several compression methods, but they do not take into account the fact that certain areas in images are of low interest to specialists, where compression methods may be used with higher compression rate and higher information loss, or even some data can be removed. Removing non-important regions makes possible to obtain a higher compression ratio and minimal loss of information. In this paper we present a compression methodology taking into account regions of interest is presented in the context of medical images, more specifically, for computer tomography of the abdominal region. Experimental evaluation shows an improvement up to 23.88% in the compression rate by using the proposed methodology.