MR Image Estimation from Sparsely Sampled Radial Scans
Ghla Stijnman, Danielle Graveron‐Demilly, Ftaw Wajer, DIRK VAN ORMONDT · 1997
This work concerns sparse radial scanning (SRS) and image estimation from the raw k-space data. We present novel reconstruction methods suitable for our own radial scan technique 'SRS-FT' and the related 'linogram'. Moreover, we omit radial trajectories to reduce the total scan time. The estimation of an image involves i) Generation of a starting image by transforming samples on each trajectory to a fixed Cartesian grid in the image domain, requiring no interpolation. ii) Removal of sparse-sampling artefacts by a Bayesian image estimation that obviates density correction. iii) Simple correction of acquisition 'dead-time'. Keywords--- Rapid MRI, sparse radial sampling, linogram, Bayesian image estimation, dead-time correction. I. Introduction MRI scans are sampled in k-space [1]. In various applications, rapid scanning is indispensable. The notion of 'rapid' may pertain either to the acquisition time of a single scan-trajectory or to the acquisition time of all scan-trajectories. W...