AUTOMATION OF CAROTID ARTERY STENOSIS MEASUREMENT IN ULTRASOUND IMAGES USING IMAGE SEGMENTATION TECHNIQUE

Abstract

Atherosclerosis is one of the foremost cause of cerebrovascular diseases like heart attack and stroke worldwide. Accurate measurement of IMT and detection of carotid artery plaques in ultrasound images is crucial in the early diagnosis of these pathologies. Manual measurement of IMT and percentage carotid artery stenosis in hospital setups are more error prone due to user interaction and speckle noise in ultrasound images that eventually degrades the image resolution leading to inaccurate measurement of risk indicators of stroke. Recently, a few image processing techniques are used to avoid shortcomings of manual measurement of IMT and carotid artery plaque by medical experts. In this study, an automated segmentation technique for carotid artery plaque from ultrasound images is presented. B-mode gray scale ultrasound images are used to detect the plaque based on the algorithm as image acquisition, feature enhancement, artery wall matching, artery width (distance) matching, artery wall selection and enhancement, bounded distance map and determining carotid artery stenosis. Experimental results on 26 longitudinal images show the accuracy of the proposed method with the aid of medical expert’s reports. Further, the proposed method does not need user interaction in all cases. The proposed technique will be further evaluated on larger database and stenosis from transverse images of carotid artery can also be measured.