CHEST X-RAY BASED DETECTION MODEL FOR PNEUMONIA IN PEDIATRIC

Abstract

Pneumonia remains a leading cause of morbidity and mortality among children, particularly in low-resource settings such as Nigeria. The accurate and timely diagnosis of pediatric pneumonia is hindered by the scarcity of skilled radiologists and diagnostic infrastructure. This study proposes a robust, efficient, and scalable classification model utilizing transfer learning to support pneumonia detection from chest X-ray (CXR) images. The model employs pre-trained convolutional neural networks (CNNs), fine-tuned under a novel Resource-Constrained Medical Transfer Learning (RCMTL) framework, to optimize predictive accuracy, computational efficiency, and equipment robustness. The approach shows promise in enhancing clinical decision-making, especially in under-resourced environments, and paves the way for practical AI integration in healthcare delivery.