Una Técnica Bayesiana y de Varianza Mínima para Segmentación del Lumen Arterial en Imágenes de Ultrasonido

Abstract

Cardiovascular diseases (CVDs) are the worldwide leading cause of deaths. Based on ultrasound, the primary assessment of CVDs is measurement of the carotid intima-media thickness and brachial endothelial function. In this work we propose im- provements to the automatic arterial lumen detection metho- dology, fundamental for the cited tests, presented in (Calderon et al., 2013); based on graphs and edge detection. We propose a bayesian approach for segmenting the minimum spanning tree of the graph created with intermediate points between edges. Lumen is located applying three criteria on segmented trajec- tories: length, dark and, our proposal, minimum variance. In 294 sonograms, mean error in brachial near wall detection was 14.6μm and standard deviation of 17.0μm. For far wall it was 15.1μm and standard deviation of 14.5μm. Our methodology maintains superior performance to results in recent literature that the original methodology presents; but surpasses it in ove- rall accuracy.