Lung Nodule Detection and Classification using 2D and 3D Convolution Neural Networks (CNNs)

Abstract

In the world, lung malignancy is one of the major causes of death related to cancer. In the initial stages, patients do not show any symptoms or signs, resulting in 75% of patients being diagnosed in a crucial stage. Consequently, there has been a call for lung cancer screening amongst at-risk populations. Accurate nodule detection is an essential process for detecting and diagnosing lung cancer in early stages from computed tomography (CT) scans images. Radiologists often use computer-aided detection (CAD) systems to receive a second opinion during the examination of images. Nodule classification is a crucial stage of the full process, which comes as the second phase in a CAD system, right after the candidate's detection. Its task is to differentiate the nodules into two categories, true nodules and false positives. This chapter's main goal was to compare different deep learning (DL) techniques based on two dimensional and three dimensional for lung nodule classification by evaluating their efficiency on common databases. The rapid development in the biomedical engineering field has made medical image analysis one of the leading areas of advanced research and development. This development and advancement's main reasons are machine learning (ML) based techniques used for medical image analysis. DL is a prominently used ML technique where a neural network is used for automatic features learning. In handcrafted feature extraction methods, the features selection, computation and calculation are challenging tasks. Instead of handcrafted feature extraction methods, deep convolutional networks (DCN) are actively used to analyze medical images, consisting of application areas, i.e., CAD, classification of disease, segmentation, etc. In this review, we have presented the state-of-the-art medical image analysis review using DCN and also highlighted the major challenges regarding this area of research.