Lossless, low-complexity, compression of three-dimensional volumetric medical images via linear prediction

3-D volumetric medical images, as for example magnetic resonance (MR) and computed tomography (CT) images, are an important source of digital data that need lossless compression to be stored or transmitted. In this paper we propose a low complexity, lossless, compression algorithm for the compression of 3-D volumetric medical images that exploits the three-dimensional nature of the data by using 3-D linear prediction. Experimental results are reported that are comparable, and in average outperform, the state-of-art results. Moreover the algorithm we present, for its low complexity, in terms of both CPU usage and memory, is suitable to be easily used also in situations in which computing power might be an issue.

I. Introduction

Nowadays there is an important demand for an efficient compression of medical images. This is motivated by the increasing importance that this data has in medical applications, such as telemedicine, tele-radiology, real time tele-consultation, PACS (Picture Archiving and Communication Systems), etc..