Liver Tumor Classification Using Pragmatic Neural Networks for Health Care Applications

Abstract

The present research involves four phases: the first phase deals with the selection of suitable filter for de-noising the liver tumor image. The performance of filters such as the Median Filter, Mean Filter, Lee Filter, Frost Filter and Fractional wavelet Filter is presented. The analysis of liver tumor image is carried out based on the performance parameters such as mean square error and Peak Signal to Noise Ratio (PSNR). Among the above mentioned filters for de-noising the image without affecting the edges, the fractional wavelet filter performs well. In the second phase, the de-noised image is segmented using rib-region partition algorithm. Features are extracted and selected using Lesion Zone feature vector algorithm. Intensive artificial neural network is used to classify the image. The classification accuracy obtained is 99.15%. Which is found to be high for the IANN method when compared with the other existing methods. newline