Automated Measurement Method Of Common Carotid Artery Intima-Media Thickness In Ultrasound Image Based On Markov Random Field Models

Cardiovascular diseases(CVDs)have been the leading threats to human healthy,common carotid artery intima-media thickness(IMT),which is considered as an important indicator for the development of cardiovascular diseases(CVDs),is widely used in clinical.Ultrasound imaging is one of the important tools for clinical diagnosis.It has the advantage of real-time,no invasion and low cost,and is considered as the preferred medical image of carotid artery.In order to solve the problems such as heavy work and high intra-inter-observer variability of manual measurement,it is necessary to consider an automatically and efficient computer-aided IMT measurement system.In this thesis,an automatic IMT measurement algorithm based on fuzzy clustering and Hidden Markov Random Field(HMRF)model is proposed.The proposed method region of interest(ROI)is extracted automatically based on prior knowledge,and the contours of the intima-media are established by fitting the edge map,and then the improved fuzzy c-means(FCM)which is based on HMRF is applied to segment ROI.The ground truth(GT)was the average manual measurements from two clinical experts with database of 80 B-mode carotid ultrasound images.The correlation coefficient between two databases reaches to 98.52%,and the mean of the absolute error between them is 0.022 ?0.0164 mm.Experimental results show that the proposed method has strong robustness against ultrasound artifacts,and achieves similar clinical parameters to GT.The proposed method could be used to provide supplementary information for the early prevention of cardiovascular disease.