Smart Imaging System for Tumor Detection in Larynx Using Radial Basis Function Networks

Tumor in Larynx is a fatal disease, which is capable to make the humans speechless. The Computed Tomography (CT) images are used to detect the benign and malignant tumors in larynx. The objective of this project is to develop an organized scheme to analyze and evaluate its probabilities with the help of a typical user friendly simulation tool like MATLAB. The innovation in this scheme has a well developed strategy for detection of malignant and benign tumors using parallel computing image related machine learning algorithms. The preliminary stage includes noise removal and feature extraction using Principal Component Analysis (PCA) from the region of interest in the larynx. This method uses images from the open source data base. The feature set extracted serves as the input for training the Radial Basis Function Network (RBFN) so as to identify the malignant and benign tumor present in the larynx. The Diagnostic Efficiency (DE) is found to be 94 to 95%.