CNN Architectures in Lung Carcinoma Nodule Identification: Detailed Analysis with Performance Comparison

Lung cancer is considered one of the deadliest diseases that have affected many lives in recent years. Early diagnosis of the disease can help in a successful prognosis. Therefore, cancer detection in the early stage is a hot research topic nowadays. CT scans play a crucial role in detecting and diagnosing cancer. They are a valuable tool in the field of medical imaging. This paper uses various deep learning models: AlexNet, VGG-16, VGG-19, MobileNet, DenseNet, and ResNet50. This study aims to find a better classification model that suits the real-time data collected for nodule detection. Better classification is achieved through better feature extraction methods. Among the analyzed models, MobileNet is considered the lightest one. At the end of this paper, a hyperparameter tuning strategy is performed to improve the overall performance of the model such that it approaches the same accuracy as in VGG-16.