AI-Enhanced Pediatric Pneumonia Classification

Pneumonia remains a leading cause of mortality among children under 5 years old, with a significant impact in low-income regions. This study explores the application of transfer learning with convolutional neural networks (CNNs) to improve the classification of pediatric chest X-rays into normal, bacterial, and viral Pneumonia. Utilizing pre-trained models on the ImageNet dataset, specifically InceptionV3, these models were repurposed for classification tasks. The framework included both binary and multi-class classification approaches, with the best binary classifier achieving an F1-score of 0.941 and an accuracy of 91.44%. The multi-class classifier reached an F1-score of 0.597 and an accuracy of 83.78%. These results demonstrate the potential of advanced AI techniques to enhance diagnostic accuracy and address challenges in Pneumonia classification, particularly in resource-limited settings.