Vascular system reconstruction from MR images using active appearance model

Abstract

Vascular system recognition and spatial reconstruction using MR images consist an important element of modern health care. The developed reconstruction method successfully handles the intensity inhomogeneity or intensity non uniformity (INU), that is an undesired phenomenon during measurement and represents the main obstacle for MR image segmentation and registration methods. The segmentation is realized by a fuzzy c-means (FCM) algorithm together with the INU estimation. The proposed method determines the contours and using a medical knowledge base analysis determines the edges that are parts of the vascular system. Finally a spatial reconstruction is performed from the obtained data. Several MR databases were analyzed, and approximately a 98.5% recognition performance was obtained. The developed method can serve as excellent support for 3-D registration and visualization techniques.