Statistical 3D Vessel Segmentation Using a Rician Distribution

This paper presents an extended version of the fully automated 3D cerebral vessel reconstruction algorithm developed by Wilson and Noble [11] which is applicable to time-of-flight (TOF) and phase contrast (PC) magnetic resonance angiography (MRA) images. We introduce a Rician distribution for background noise modelling and use a modified EM (Expectation-Maximization) algorithm for the parameter estimation procedure. The proposed algorithm is applied to PC-MRA images. It is shown that the estimated Rician distribution gives a better quality-of-fit to the observed background noise distribution than a Gaussian distribution. In the experiments reported, the segmented 3D vasculature is shown to be qualitatively comparable with the results obtained from higher resolution TOF MRA images.