Detection System for Lung Tumours Based on CT Scan Images Using Cellular Automata

Lung tumours are masses of tissue that arise abnormally in the lung organ. In the field of medicine, One means of detecting and analyzing lung tumours is to use Computed Tomography (CT) scan images, where doctors see with the naked eye on the image and make a diagnosis. The existence of a lung tumour area detection system using a computer program is needed as a companion system in diagnosing. This research develops a lung tumour detection system using Cellular Automata (CA) with the system input is CT scan image. The quality of CT Scan image is improved through pre-processing, then segmentation is performed to separate the lung object from the background. Binaryization is applied to the segmented image to clarify the presence of lung tumour area. The detection process by CA is conducted on the binaryized image. The experiment was implemented on 20 CT scan images for normal lung and 20 other images for abnormal lung obtained from The Lung Image Database Concortium (LIDC) and Image Database Resource Initiative (IDRI). Analyzing the experimental data using performance score, it was found that the accuracy of detection was 95%, 100% precision, 90% sensitivity, and 100% specificity. These conditions indicate that the detection system has good reliability.