Lung Pattern Classification for Interstitial Lung Diseases Using a Deep Convolutional Neural Network

Abstract

Lung nodules is a circle or oval that is often found on chest radiograph (X-ray) or computed tomography (CT) scan. Lung nodules usually do not cause a symptom, usually found incidentally on chest Xray or CT. Diagnosis process performed by observing CT Scan. The further process by taking a sample of tissue in the lungs and perform laboratory tests. This procedure is not effective. The aim of this study was develop software to detect lumps on pulmonary CT scan images are nodules or artery. This study consists of five steps. The first step is to build user interface. The second is to develop software for lung segmentation using Active Shape Model. The third is to develop software for candidate pulmonary nodules segmentation using mathematical morphology. The fourth is detection of pulmonary nodules using Support Vector Machine (SVM). The fifth is a 3-D visualization of pulmonary nodules using volume rendering.