Ensemble of CNN and DenseNet121 Models for Pneumonia Prediction Using X-Ray Images

When left untreated, pneumonia, a respiratory infection, can result in serious sickness or even death. In recent years, promising findings have been obtained in the accurate identification of pneumonia utilising X-ray image processing methods based on machine learning. In this article, we provide a deep learning-based approach for predicting pneumonia from X-ray images. We applied the averaging technique to combine CNN and DenseNet121, two well-known deep learning models, using a dataset from Kaggle. According to our tests, the CNN model had a test accuracy of 91.2% whereas DenseNet121 had an accuracy of 89.1%. The ensembled model's accuracy of 90.2 % shows how assembling can improve the accuracy of individual models. Our findings imply that deep learning models may prove to be a valuable tool for the precise and early detection of pneumonia, which may lead to better patient outcomes.