Автоматическая сегментация эпикардиально жировой ткани и количественная оценка радиомических параметров при компьютерной томографии сердца

Abstract

This paper presents a computerized tool for automatic segmentation of epicardial adipose tissue (EAT) on CT heart images, aiding in cardiovascular disease diagnosis. Manual quantification methods suffer from time consumption and variability. Our tool combines automated segmentation with manual adjustments, enhancing accuracy and efficiency. It utilizes basic image processing, ensuring computational simplicity. The proposed algorithm improves diagnostic support, offering potential for personalized therapy. Additionally, the work includes analysis of Region of Interest (ROI) using radiomic analysis, enabling texture analysis of images to further study EAT. The paper also features a segmentation algorithm and a prototype application integrating semi-automatic segmentation mechanisms.