CAD System for Non Small Cell Lung Carcinoma using Laws’ Mask Analysis

Abstract

Lung carcinoma is most occurring death through cancer across the world. For diagnosis and detection of lung cancer there are different techniques used. The most encouraging techniques for early detection of cancerous cells are Computer Aided Diagnosis (CAD). CAD depends on the analysis of quality of ultrasonic images by detecting lesions that may indicate the presence of lung cancer. The CAD system envelopes four main processing steps: preprocessing, feature selection, feature extraction, and feature classification. Different classifiers are used to allocate the cells into adenocarcinomas, squamous cell carcinomas and large cell carcinomas which are the parts of Non Small Cell Lung Carcinoma. This paper proposes a CAD system using Laws’ mask and SVM classifier. The accuracy of 95.65% is obtained from laws’ mask 3. The results will be further used for a CAD system for early analysis of lung cancer to improve the chances of survival of patient.