Pulmonary Nodules Binary Classification using CNN and LSTM

Pulmonary cancer is amongst the major reasons of death all across the globe. The rate of survival in patients suffering from lung cancer increases only if it is diagnosed at early stages. Nowadays, the employment of Computer-Aided Detection (CAD) systems can lead to the rapid detection of lung cancer and potentially save lives. Deep learning has been arising as an optimistic area of artificial intelligence, and this has attracted rapidly increasing recognition in the medical field. A classification model for lung nodules that utilizes Computed Tomography (CT) scans and relies upon a Convolutional Neural Network (CNN). Images of lung nodules extracted from the CT scan images available in the LIDC-IDRI dataset are utilized by the model. The extraction of nodules has been done utilizing the python library package named pylidc. The graphs have been compiled using tensorBoard. The maximum accuracy attained using the proposed model is 83.86% and the loss was 32.81%.