An Improved Statistical Approach for Cerebrovascular Tree Extraction.

In this paper, we present a statistical approach to aggregating shape and speed information for whole cerebrovascular tree extraction in time-of-flight magnetic resonance angiography (TOF-MRA). By embedding Frangi's vesselness measure into the prior mopodel, the newly porposede segmentation framework can greatly improve the capability of detecting the tiny vessel branch. Keywords: Statistical segmentation, Time-of-flight, Hessian matrices, maximum a posteriori(MAP) estimation, Markov Random field. [ABSTRACT FROM AUTHOR]