AI-Enhanced Diagnosis: Pediatric Chest X-ray Classification for Bronchiolitis and Pneumonia

The pediatric diseases in question are bronchiolitis and pneumonia, which pose a significant threat to children, especially those under ten years of age. Rapid diagnosis often requires a chest X-ray; reading and interpreting these images is challenging, and requires the expertise of a skilled doctor. It is essential to take advantage of advanced image recognition techniques to aid in interpreting these examinations and extracting necessary information. This study employed deep transfer learning models, including VGG16, VGG19, MobileNetV2, and InceptionResNetV2, to diagnose bronchiolitis and pneumonia in pediatric chest X-rays (PCXr) for the first time. Our findings show that the InceptionResNetV2 model has achieved the highest recall rate for bronchiolitis, with an impressive value of 78.82%. Following that, VGG16 achieved a recall rate of 77.64%, MobileNetV2 at 74.11%, and VGG19 at 62.35%. Furthermore, when assessing the models comprehensively based on their performance in terms of the F-score, InceptionResNetV2 outperformed the others with an F-score of 65.68%.