Convolution Neural Network: A Proposed Method for Detecting Pneumonia from X-ray Images

Pneumonia requires quick and accurate diagnosis for successful treatment as it is causing morbidity and mortality all around the world. Deep learning, especially convolutional neural networks (CNNs), has lately evolved to show interesting potential for automated pneumonia diagnosis and categorizing chest X-ray images. With an increasing diagnostic accuracy and support for clinical decision-making, this study investigates CNN's use in Pneumonia classification. The research is based on a CNN model using a vast collection of labelled chest X-ray pictures containing both normal and pneumonia-positive individuals. Our CNN's architecture consists of several convolutional, pooling, and totally linked layers meant to automatically extract and learn essential features suggestive of Pneumonia. Performance parameters including Accuracy, recall, F1-scoreand Precision evaluate the performance of model. Using a separate validation dataset of 5856 photos, the model's accuracy was evaluated. The outcomes of the investigation showed that the model's general accuracy was 91%. By use of individual X-ray image processing, the suggested method has several possible uses in clinical practice, including facilitating doctors in precisely classifying Pneumonia severity levels.