A Wavelet Fusion Approach to the Reconstruction of Isotropic-Resolution MR Images from Anisotropic Orthogonal Scans

Iman Aganj¹, Christophe Lenglet², Essa Yacoub², Guillermo Sapiro¹, Noam Harel²

Hardware, timing, and SNR considerations restrict the slice-selection and the in-plane resolutions differently, generally resulting in thicker acquisition slices and therefore anisotropic voxels. This non-uniform sampling can be problematic, especially in image segmentation and clinical examination, since the image will be missing high frequencies in the slice-selection direction. High-resolution MR volumes with isotropic voxels are acquired at the cost of requiring the subject to be motionless for a clinically unreasonably long time, hence increasing the risk of motion artifacts. This can be alleviated by dividing the acquisition into (two or) three separate scans, with thicker slices yet complementary resolutions. Every scan will then have a shorter acquisition time and a lower chance of undergoing motion-related distortion. Misalignment between the three scans can be corrected by employing a variety of available registration techniques, and the high-resolution image is eventually reconstructed from the three scans. In this work, we adopt a non-iterative wavelet-based approach, which takes into account the actual system response of the MR scanner. We show results from three orthogonal Susceptibility-Weighted Imaging datasets acquired at 7T, and compare them with a high resolution ground truth dataset.