Adaptive fuzzy clustering-based texture analysis for classifying liver cancer in abdominal CT images

Segmentation of diseased liver in abdominal CT images is a challenging task due to variations in shapes, tissue similarity between adjoining organs. We propose an automatic detection technique that integrates the fuzzy clustering with adaptive thresholding for segmenting the liver and finding the tumour region in abdominal CT images. Various features like texture features, morphological features and statistical features have been extracted from the output images and used as input to the neural network classifier to classify the malignant and benign tumour of the liver. The method was evaluated in a series of 45 images collected from medical image computing and computer assisted intervention (MICCAI) database and the efficiency is tested in terms of sensitivity, specificity, and accuracy. We obtained the accuracy of 97.82%, 95.74% in BPN and LVQ and higher accuracy of 98.82% is achieved with PNN in detecting tumours which are comparable to published results.