Carotid Ultrasound Plaque Classification Using a Combination of Region of Interest, Discrete Wavelet Transform and Neural Networks

Carotid ultrasound plaque classification using a combination of .region of interest, discrete wavelet transform and neural networks are used to classify the symptomatic or asymptomatic ultrasound plaque images. The system involves three steps: 1) the preprocessing step using the ROI technique. 2) The method Feature extraction is done by averaging values and discrete wavelet transform.3) Classification using a Neural Network. In this work Multilayer feed-forward neural network is adapted for training and testing the images. The accuracy of 70% is obtained in proposed work using neural network classifier.