A deep-based compound model for lung cancer detection

Abstract

$X$-ray image analysis is primarily performed by medical specialists. Patients expect a correct interpretation of these images regardless of cost. Despite various advantages of chest radiography, the interpretation of Magnetic Resonance Imaging (MRI) has always been a major issue for the physician and the radiologist due to misdiagnosis. According to the World Health Organization, Lung cancer cost around 1.8 million deaths in 2020, which makes it the leading cause of cancer death worldwide. Late diagnosis and lack of means of screening are the main problems. The algorithm can help radiologists accurately estimate the malignancy risk of lung nodules. This paper aims to detect and classify lung cancer using deep learning. We used the Convolutional Neural Network (CNN) algorithm combined with the Faster Regions with CNN (Fast R-CNN). Our model provides very encouraging results compared to those obtained by the work of the literature, which provides a model with a high accuracy rate for medical assistance.