Pneumonia in Children Using Artificial Intelligence and Machine Learning

Pneumonia is an important cause of childhood morbidity and mortality globally, therefore accurate and timely identification is vital. While chest X-ray imaging is still the standard diagnostic method, manual interpretation is haphazard and subjective. In recent years, advances in Artificial Intelligence (AI) and Machine Learning (ML) have allowed for systematic and automated processing frameworks that now show improved performance. This work investigates the comparison of four models, including: a Convolutional Neural Network (CNN), an Artificial Neural Network (ANN), a model from the Gradient Boosted (GBoost) family, and a CNN with mobile net architecture (MobileNetV2) based on transfer learning using a Kaggle pediatric chest X-ray dataset containing 5,800 images labelled for pneumonia or normal. MobileNetV2 achieved the best overall classification accuracy at 90.06%, while CNN achieved 87.18%, ANN 84.46%, and GBoost 74.52%. This work demonstrates that lightweight software architectures based on efficient mobile deep learning models such as MobileNetV2 can continue to be promising in the case of pediatric pneumonia classification.

Перевод пока недоступен