The carotid intimal-media thickness measurement of Ultrasound images based on an improved Chan-Vese model

Stroke is one of leading threats to human health. The intima-media thickness (IMT) of common carotid artery (CCA) is one of major criteria for early screening and diagnosis of stroke. Ultrasound imaging is widely used for the intima-media thickness measurement because it is noninvasive, economic and safe. So far, physicians manually measure the IMT by using Ultrasound imaging, and the procedure is subjective, time consuming and unreliable. In this paper we present a semiautomatic method to measure the carotid intima-media thickness based on Ultrasound images. The initial procedure is to select a region of interest (ROI) in the concerned carotid ultrasound image. A modified level set method is employed to segment the IMT by adding local intensity information to the variational Level Set formulation. This mechanism is used twice. In the first step, the boundary between lumen and intima (LI) can be segmented from ultrasonic images by the modified Chan-Vese method. After we get the LI boundary, we use the similar method to detect the boundary between the media and adventitia (MA). After LI and MA are detected, the IMT can be easily calculated. We compare the result of our method with other methods, and the results show that our method can measure the IMT accurately with less manual inputs and without any preprocessing steps.