On the lossless compression of multidimensional medical imagery

Nowadays, many technologies involved for medical examinations produce multidimensional images. Such data need to be managed in an effective manner in order to be efficiently stored and transmitted. In such scenarios, data compression techniques are essential to improve the efficiency of transmission and storage. Lossless compression techniques are generally preferred, since medical images are often sensitive and important data. Indeed, through lossless compression techniques, the original data can be exactly restored. In this paper, we define a predictive structure well suited for the lossless compression of multidimensional medical images. We experimentally tested our approach on several datasets, including a dataset of 3-D Computed Tomography (CT), 3-D Magnetic Resonance (MR) and 5-D fMRI images. The experimental results we achieved outperform other state-of-the-art approaches for 3-D medical images.