Fuzzy Quantification of Artery Lesions in Renal Arteriographies

This article presents a computer-based system for the automatic quantification of vascular lesions from 2D renal angiograms. The different stages of the system are described. The quantification relies principally on the segmentation process that provides the outlines of the arteries. The segmentation method is based on the cooperation of two fuzzy segmentation operators in order to take into account the natural stepped process of an expert that looks for useful information in the image. The first operator is based on a linguistic description of an edge and guided by the morphological attributes of the aorta. The second operator deals with a fuzzy clustering based on the fuzzy-C-Means algorithm to detect the artery boundaries. Beyond the segmentation process, a fuzzy automaton is defined to recognize the lesions along the artery. The procedure is based on the estimation of the reference healthy artery which is represented within the context of the fuzzy set theory. A quantification procedure follows the syntactic analysis and provides fuzzy results which are reproducible measurements of artery lesions. An important objective of the present work was to test the performances of the designed system on the degree of stenosis. For 49 images, both the radiologist caring for the patient and the system were compared to a gold standard. Compared to individual radiologists, the computer system gave a more precise estimation of percent stenosis and did not over or under estimate the severity of the lesion.