Performance Analysis of Hyperparameters of Convolutional Neural Networks for COVID-19 X-ray Image Classification

Analysis of chest X-ray images of COVID infected patients is one of the important diagnostic strategies. The manual identification of these images may be erroneous and faulty. So the computer-aided diagnosis of COVID infections using deep learning techniques is becoming useful. In this paper, the classification of chest X-ray images using CNN is conducted, and the performance of different optimizers is studied. The dataset containing chest X-ray images of normal and COVID infected patients is collected from Kaggle. The experimental study suggested that Adam optimizer achieved 95.83% classification accuracy, and it outperformed the other three optimizers.