Pulsative Flow Segmentation in MRA Image Series by AR Modeling and EM Algorithm
Segmentation of CSF and pulsative blood flow, based on a single phase contrast MRA (PC-MRA) image can lead to imperfect classifications. In this paper, we present a novel automated flow segmentation method by using PC-MRA image series. The intensity time series of each pixel is modeled as an autoregressive (AR) process and features including the Linear Prediction Coefficients (LPC), covariance matrix of LPC and variance of prediction error are extracted from each profile. Bayesian classification of the feature space is then achieved using a non-Gaussian likelihood probability function and unknown parameters of the likelihood function are estimated by a generalized Expectation-Maximization (EM) algorithm. The efficiency of the method evaluated on both synthetic and real retrospective gated PC-MRA images indicate that robust segmentation of CSF and vessels can be achieved by using this method.
KeywordsInitial Segmentation Linear Prediction Coefficient Acardiac Twin Optimal Feature Selection Intensity Time Series
Unable to display preview. Download preview PDF.