A Comparative Study For Pneumonia Diagnosis Useing the Deep Learning Techniques

Abstract

Pneumonia, a respiratory infection that causes inflammation of the air sacs in one or both lungs, is a significant contributor to mortality worldwide. Timely detection of acute pneumonia is crucial as it can be fatal without prompt medical intervention. Radiographic imaging, specifically X-ray imaging, is commonly used for diagnosing pneumonia. This study proposes a technique based on deep learning algorithms, including Convolutional Neural Network (CNN), DenseNet-121, and VGG-16, for analyzing X-ray images to detect pulmonary inflammatory diseases. This study was trained on a database comprising 5,216 X-ray images. The results of the study demonstrate that the CNN algorithm outperforms the DenseNet-121 and VGG-16 algorithms in terms of accuracy. Consequently, a program employing the CNN algorithm was developed to assist healthcare professionals in the diagnosis of pneumonia.