Lossless and near-lossless motion-compensated 4D medical image compression

Four-dimensional (4D) medical images, which are sequences of volume images over time, are represented by a great volume of information. One set of 4D medical images can easily take up hundreds of megabytes of storage volume. In this paper, we present a technique for compressing 4D medical images by combining motion compensation and lossless/near-lossless image encoding. Our technique incorporates a new fast 3D cube match algorithm that effectively exploits the redundancy between frames. The proposed scheme is validated by experiments on 4D cardiac images.