Segmentation of Doppler Carotid Ultrasound Image using Morphological Method and Classification by Neural Network

In the recent times, image segmentation plays a critical role in medical study for taking individual decisions by physicians. This technique attempts to estimate the structure of lumen and plague segmentation on the transverse view of B-mode ultrasound images of common carotid artery (CCA). The proposed method segments both the lumen and plague, whereas only the lumen is segmented in the conventional methods. The lumen contours are segmented using self-adaptive histogram equalization, nonlinear filtering, sobel edge detector and morphology methods. Plaque in the blood vessel is examined by automatic detection using Fuzzy C-Means (FCM) filtering and canny edge detector methods. The significant benefit of the proposed method has been numerically validated on real time image data from RCT‘S (Philips HDI 5000 Ultrasound Scanner) and results show that the proposed method performs in a better way when compared to existing techniques. The proposed method is simulated using MATLAB (2013a). Keywords— Anisotrophic filter, Segmentation, Lumen, Plague, FCM filtering , Classification, Neural Networks.