Study Of Carotid Artery Segmentation Based On Ellipse Fitting And Improved FCM

As the blood vessels segmentation technique is widely used in clinical medicine, our expectation for precision of carotid artery segmentation and accurate positioning of patches is higher and higher. Blood vessels segmentation techniques are mainly used for helping the doctor to analyze and judge the probability of certain diseases. We can clearly observe whether vascular shape changes and there are patches in the blood vessels through the segmentation of carotid blood vessels and help doctors monitor the narrow degree of blood vessels in real time.Papers presented carotid artery segmentation method has higher accuracy.There is a low gray and isolated noise phenomenon in our initial medical image,By using piecewise linear transformation to stretch the image, image filtering in the use of Gauss function. At the end of the image using the horizontal and vertical direction of the two gray integral projection algorithm the positioning location of vascular. For similar elliptical shape in the outer wall of the blood vessel image presented, we fit the ellipse by calculating the least-squares algorithm error distance function. On this basis, the change of the elliptic long,short axis and angle, and choose the best ellipse uses laplasse operator condition gray-scale image. But the elliptical selected and real wall boundary there is a big gap, so we will be adjacent pixel gray scale information is added to the cost function of dynamic programming, the ellipse is optimized. After the end of the outer wall of the blood vessel segmentation, the inner boundary is not very clear,and the surrounding tissue pixel difference is not large, we add a constructor G on the vascular lumen segmentation in the objective function of fuzzy C means algorithm. The improved FCM algorithm can effectively avoid the blood vessel lumen segmentation process in local area, get the global optimal solution, and also optimize the lumen of the segmented boundary. At last, we improved FCM algorithm in the lumen of plaque segmentation, to determine the location ofplaque by calculating the lumen and plaque proportion.In this paper based on the MATLAB 2012 b platform, using the improved method for image segmentation. The segmentation results and the manual segmentations(gold standard) were compared. In the experiment, the average value of AO reached 0.952, verify the effectiveness of the algorithm. The improved algorithm and the general algorithm of image segmentation, by comparing the results, the common method to get the average value of AO is0.941.According to the trend of AO values determine the specific location of plaque, our algorithm has good effect on the extraction of carotid artery.