Detection of Intima-Media Layer of Common Carotid Artery with Dynamic Programming Based Active Contour Model

Ultrasound measurements of the carotid artery wall in image are usually obtained by manually tracing. In this paper, we present an automatic segmentation method to detect the intima-media layer in far wall of the common carotid artery. The energy definition of active contour model is used. Different from the traditional approach applied in snake techniques, we treat the optimization problem as finding the shortest cost path in a directed graph. Dynamic programming is selected to search the shortest path. The external force and internal force in snake model are modified to be suitable for our approach. To reduce the effect of speckle noise, a new method in speckle reduction by anisotropic diffusion is adopted. At last, we compare the result of our method with other two methods. Results show that our method can detect the intimal and adventitia layers as well as other methods, the two layers will not cross with each other as traditional dynamic programming does. Moreover, our method needs less manual input than others.