Image Classification and Clustering using Wavelet Based Radial Basis Neural Network
P. Sathish · Zenodo (CERN European Organization for Nuclear Research) · 2016
Accurate image segmentation and classification, is essential for medical diagnosis of scans. Of late, magnetic resonance (MR) images have become the commonest tool of clinical investigation. In this study we address to clarify brain tumor images into normal, non cancerous (benign) brain tumor and cancerous (malignant) brain tumor by collecting the complete history of the patients in terms of their food habit, life style and the severity of the age. The proposed method follows three steps, (1) wavelet decomposition, (2) texture feature extraction and (3) classification. Discrete Wavelet Transform is first employed using Daubechies wavelet (db4), for decomposing the MR image into different levels of approximate and detailed coefficients and then the gray level co-occurrence matrix is formed, from which the texture statistics such as energy, contrast, correlation, homogeneity and entropy are obtained. The results of co-occurrence matrices are then fed into a radial basis neural network for further classification and clustering for tumor detection. The proposed method will be applied on real MR images, and on all types cancer images and the accuracy of classification using radial basis neural network will be rigorously evaluated and the physician can diagnose and design the better therapies.