Multi-Classification of Lung Infections Using Improved Stacking Convolution Neural Network

Lung disease is a respiratory that poses high risk to people worldwide and includes pneumonia COVID-19. As such, quick precise identification of lung vital in medical treatment. Early detection diagnosis can significantly reduce the life-threatening nature diseases improve quality life human beings. Chest X-ray computed tomography (CT) scan images are currently best techniques detect diagnose infection. The increase chest or CT at time training addresses overfitting dilemma, multi-class classification will deal with meaningful information overfitting. Overfitting deteriorates performance model gives inaccurate results. This study reduces issue computational complexity by proposing new enhanced kernel convolution function. Alongside an function, this used neural network (CNN) models determine Each CNN was applied collected dataset extract features later these as input models. shows extracting deep from common layers increased procedure. improves diagnostic performance, evaluation metrics improved support vector machine (SVM). results were obtained using SVM classifier fed provided CNN, success rate 99.8%.