Hybrid Neural Network Method Using Aquila Optimization Algorithm for Detection of Lung Cancer in CT Images

The existence of a lung nodule implies the existence of lung cancer. The classification of lung nodules as benign or malignant has frequently been done using deep convolutional neural networks (DCNNs). The five-year rate of survival of lung cancer must be increased with early pulmonary nodule diagnosis. To aid radiologists in diagnosis, numerous computer-aided diagnosing (CAD) techniques have been developed for nodule detection. Deep convolutional neural networks (CNNs) are typically utilized to detect regions of the primary pulmonary nodules without taking into account nodule’s surrounding tissues. In this work, a semi-supervised neural network is developed for early detection of pulmonary nodules with global optimization using Aquila optimizer (AO) performed on LIDC-IDRI dataset. The proposed model delivers qualitative results on comparing with existing methods achieving 98.23% of accuracy, 98.1% of sensitivity, 97.67% of specificity and 97.9% of precision.