Investigation on deep learning based optimization of convolutional neural network for lung tumor identification in big data

Abstract

In many countries, lung tumor is the most threatening disease of newlinehuman bodies because of the rising death rate. The Computed Tomography newline(CT) scan plays the major role in staging and diagnosis of lung tumor. newlineTherefore, CT can identify calcification which is a crucial sign of a benign newlinetumor. But if the tumor consists of soft tissue attenuation, then the diagnosis newlineof tumor will become a challenging one. In each year, the death rate caused newlineby lung tumor is increased by 0.5%. Hence early detection of tumor and newlineproper treatment is necessary for tumor patients. newlineA Computed Tomography (CT) scan is a medical imaging approach newlineused in the thoracic radiology field and also it helps to detect the lung tumor newlinein earlier stage. In general, the lung tumor is divided into two types like newlinemalignant tumor and benign tumor. The malignant tumor spreads the entire newlinehuman body and affects other parts with its growth. The benign tumor is not newlinethat much dangerous than malignant. The beginning stage of the lung cancer newlineis termed as primary lung cancer which is also called as carcinomas. The newlineprimary type of tumor is classified as Small Cell Lung Carcinoma (SCLC) newlineand Non-Small Cell Carcinoma (NSCLC). In the advanced technology, newlinevarious techniques are established for the detection of lung tumor. Due to newlinevarious critical factors, the recognition of lung tumor is the major challenge in newlinemedical field. Recently various deep learning techniques are developed for newlinelung tumor detection in the beginning stage. These effective techniques detect newlinethe tumor accurately and also assist the patients by decreasing the mortality newlinerate due to lung tumor. In recent decades, big data plays an active research newlinedue to its various ranges of applications. The researchers explained that big newlinedata was improving in three dimensions such as variety, velocity and volume. newline