In this paper we propose efficient method for automatic estimation of intima and media boundary of common carotid artery (CCA) based on USG image analysis. The distance between intima and media is called intima-media thickness (IMT) and is used to predict cardiovascular events like myocardial infarction and stroke. At present, it is measured by manual segmentation of USG image by physician. Automatic estimation of IMT by freeing physician from this tedious work should speed the process and assure higher reproducibility. The proposed algorithm involves USG image denoising by nonlinear diffusion filtering and iterative image segmentation by modified active contour method with suitable energy function.