To address the challenge of low resolution and limited number of slices in in-vivo cDTI, and the need for 3D structural information in image-based modeling, we compare five interpolation techniques of predominant cardiomyocyte orientation: two low-rank models, one rule-based method and two tensor interpolation approaches. The direct tensor interpolation approaches result in the smallest errors, followed by the low-rank models and the rule-based method. In view of an optimal experimental design in-vivo, the ex-vivo experiments suggest a larger benefit of increasing in-plane resolution rather than SNR, and a non-linear increase in error for less than five short-axis slices.
This abstract and the presentation materials are available to members only; a login is required.