Compression of Diffusion Weighted Magnetic Resonance Images Using JPEG2000

Abstract

Magnetic Resonance Imaging (MRI) is a medical imaging technique which is routinely used for anatomical and functional imaging. Diffusion Tensor Imaging (DTI) characterizes the diffusion of water molecules using a tensor and allows computation of quantitative metrics, such as fractional anisotropy, which measures anisotropy of the diffusion process for each voxel. The image datasets used in DTI are four-dimensional with three spatial dimensions and a fourth dimension representing the direction of diffusion sensitivity applied during acquisition. In this work,we investigate the use of three-dimensional JPEG2000 for lossy and lossless compression, as well as the trade-off between the number of diffusion directions and bits/pixel used to compress each volume. These experiments show that diffusion weighted MR images can be reliably compressed using JPEG2000.