CAPSU: A completely automated method for carotid plaques segmentation in ultrasound images

The accurate carotid plaques segmentation (CPs-S) is essential in the ultrasound (US)-guided carotid atherosclerosis diagnosis. In this work, we propose CAPSU: a completely automated method for CPs-S in US images. The core of the method is an innovative initialization procedure that exploits carotid wall motion analysis to automatically initialize a level-set segmentation algorithm. A strain analysis is also used to improve the initialization procedure. CAPSU performance with and without the strain analysis were evaluated on US data from 8 patients and compared with manual contouring from an expert sonographer. Results show the CAPSU effectiveness as accurate and reliable tool for fully automated CPs-S in US images.