Classification of Bacterial and Viral Pneumonia Using Chest X-Ray Images with Convolutional Neural Network

This study examines the effectiveness of convolutional neural network (CNN) architectures, specifically Xception, InceptionV3, ResNet50V2, and DenseNet201, in classifying bacterial and viral pneumonia using chest X-ray images. By applying RMSprop and SGD optimizers, the models' accuracy and processing times were evaluated to determine the most suitable architecture for pneumonia detection. Results indicate that ResNet50V2, when paired with the RMSprop optimizer, achieves the highest classification accuracy at 0.83 and the fastest processing time at 2460 seconds. These findings suggest that the ResNet50V2 architecture, optimized with RMSprop, offers a reliable approach for distinguishing bacterial from viral pneumonia, highlighting its potential application in clinical diagnosis.