This paper describes a new method for the three-dimensional (3-D) tracking and the quantification of blood vessels from Magnetic Resonance Angiography (MRA). The approach is based on the 3D geometrical moments and consists of the following steps : (1) interactive selection of 3-D seed points ; (2) automatic tracking of the vessels ; (3) local computation of both diameter and orientation ; (4) rendering of the vessels. This detection and estimation scheme has been validated on simulated and real data.

No abstract

Textural features are compared for the classification of MR muscle images . The objective is to determine which features optimize classification rate using small ROI . Four classes of textural features are considered, we have studied fractal, cooccurrence, Higher Order Statistics (HOS) and mathematical morphology . The quantitative evaluation of the discrimination power of the features is based on the performance of the classification error rate with a K-nearest neighbor classificr . The resultsshows that the mathematical morphology features provide the best classification rate on our clinical MR images of health and ill muscles.