An Enhanced Convolutional Neural Network with Principal Component Analysis for Pneumonia Diagnosis Classification from Medical Images

Pneumonia is a lung infection that causes inflammation in the air sacs and is one of the leading causes of death worldwide for children under the age of five. The increasing prevalence of pneumonia, coupled with the critical need for accurate diagnostic tools, drives the development of advanced Machine Learning models. Therefore, this study presents an enhanced convolutional neural network (CNN) integrated with Principal Component Analysis (PCA) to improve pneumonia diagnosis from medical images. The proposed model influences the strengths of CNNs in feature extraction while employing PCA for dimensionality reduction, thus optimizing computational efficiency and reducing overfitting. We conducted experiments on a pneumonia dataset of chest X-ray images, implementing a multi-layered CNN architecture augmented with PCA to preprocess the input data. Performance metrics for evaluation were systematically evaluated against the CNN baseline model. The results demonstrate significant improvements in classification performance with an accuracy of 98%, highlighting the effectiveness of combining enhanced CNNs with PCA. This approach not only enhances diagnostic accuracy but also facilitates quicker and more efficient analysis of medical images, potentially aiding radiologists in clinical decision-making. The findings suggest that integrating traditional machine learning techniques like PCA with modern deep learning frameworks can yield robust solutions for complex medical imaging challenges.