Lung Cancer and Pulmonary Node detection using Deep Learning: A survey

Abstract

Lung cancer can present as pulmonary nodules in the early stages; however, CT screening has been shown to be useful in detecting these lesions and reducing mortality. However, because there are so few radiologists and they are already overworked, the exponential growth of picture data makes accurate appraisal of them an extremely difficult assignment. To automatically identify and categorize pulmonary nodules in medical images, a variety of techniques, particularly those based on deep learning with convolutional neural network (CNN), have been developed recently. We provide a thorough examination of these strategies’ effectiveness in this paper. First, we quickly discuss CNN’s essential concepts and the arguments for reasons why they are suitable for the interpretation of medical imagery. The environmental setup required to make lung nodule investigations with CNNs possible is then described, along with a brief overview of the relevant medical image datasets. Addition-ally, thorough analyses of current developments in lung nodule analysis using CNNs are offered. Finally, current issues and bright future options for increasing CNN’s use in medical image processing, and specifically in pulmonary nodule assessment, are discussed. It has been demonstrated that CNNs have significantly changed lung cancer early diagnosis and treatment.