Interpolation of human cardiac DT-MRI data using polar coordinates

Diffusion tensor magnetic resonance imaging (DT-MRI, or DTI) is a new technique allowing ex-vivo and/or in-vivo investigation of human fibrous tissues. Processing of DT-MRI data appears as a new field for image processing community. This paper presents an interpolation method for the eigenvector fields in human cardiac DT-MRI. The method accounts for inherent relationships between vector components and the influence of the heart shape on eigenvector fields. It consists of converting the ventricle onto a polar coordinate space, and achieving simultaneously interpolation and denoising. The interest of the method is demonstrated with the aid of human cardiac DT-MRI data.