Classification of Ultrasound Liver Image Using NLM

Abstract

The conception behind medical image classification is to categorize the ultrasound images for
medical diagnosis purpose. Tissue classification with ultrasound has become very important
nowadays for the analysis of ultrasound images. In this paper, a liver ultrasound image classification
method has been proposed for diagnosis of liver disease. The diagnosis system includes three steps:
Preprocessing, feature extraction and classification. Preprocessing of liver ultrasound images is
performed by Non-Local means filter. Fuzzy c-mean clustering method with threshold Segmentation is
applied on preprocessed image for segmentation purpose. For appropriate classification, proper
selection of features of the image is required. By applying principal component analysis features are
extracted and selected. Then ANFIS classification technique is implemented for classifying the liver
ultrasound image as Heterogeneous, Normal Fatty & Abnormal Fatty liver image. A precise
classification rate of 96% is achieved using ANFIS.