ENHANCING PNEUMONIA DETECTION FROM CHEST X-RAY IMAGES USING RESNET-18 DEEP LEARNING MODEL
Keywords:
Pneumonia Detection, Chest X-ray, Convolutional Neural Networks (CNN), Deep Learning, Digital Image Preprocessing, ResNet-18, Transfer Learning, Medical Imaging, HealthcareAbstract
Pneumonia remains a significant global health challenge, necessitating prompt and accurate diagnosis to mitigate morbidity and mortality rates. This research proposes the ResNet-18 convolutional neural network architecture for pneumonia detection from chest X-ray images. The chest X-ray images dataset comprises 5863 samples across both categories (pneumonia and normal) and was obtained from Kaggle. The proposed research compares its results against existing models such as CheXNet, VGG-19, and CNN ensembles. The experimental results show that, after the preprocessing, the model achieved an accuracy of 98%, a precision of 98.24%, a recall of 97.92%, and an F1-score of 98.08%. The proposed system has been deployed as a Streamlit-based web application, facilitating real-time diagnostics of pneumonia detection and highlighting the potential of artificial intelligence in medical image analysis.
