An Optimized Lung Cancer Classification System for Computed Tomography Images

Amongst diverse cancers, lung cancer is measured to be the foremost reason of cancer demise with utmost demise pace. Nodules lying on lungs have distinct structures, they could be either circle or coil shaped which under various circumstances composes the recognition complex. In this work a system has been urbanized for detection of lung cancer in its early stages and classification between malignant and benign tumors via images from Computerized Tomography (CT) scanner. Lung cancer detection process has four steps which include pre-processing phase, segmentation, feature extraction and lung cancer cell classification. BAT Algorithm is applied to provide considerable optimization results which improve the performance of system. The classification between malignant nodules and benign has been done through Artificial Neural Network Ensemble to provide results of higher accuracy. The overall accuracy, sensitivity and specificity of 98.5%, 100% and 91% respectively is acquired in the system.