An Improved Grey Wolf Optimization–Based Convolutional Neural Network for the Segmentation of COVID-19 Lungs–Infected Parts

The coronavirus outbreak is a recent pandemic that destroyed most of the lives, economy, and livelihoods. The detection of
COVID-19 is the main aim to detect and provide better treatment for patients to mitigate its impact. In addition, it is necessary to diagnose the disease swiftly with upgraded technologies. This can be achieved by CT image scanning. This provides
the fastest detection of the disease. Moreover, it can also be used to diagnose the percentage of the affected lung areas. To
perform this fastly, we propose a novel approach known as Convolutional Neural Network (CNN)–based Improved Grey
Wolf Optimization (IGWO) algorithm. The proposed CNN utilizes a SegNet-based approach which can be used to detect
the affected area in the lungs by using the encoder and decoder steps. The encoder in this approach uses three types of CNN
architecture. First, the decoder is used to reconstruct the images. The overfitting issues during the iterations and complexities are reduced by adopting the IGWO approach. The experimental analysis depicts that the proposed approach effectively segments the CT images and promptly diagnoses the affected lung area.