Convolutional neural network-based automated pediatric bone age assessment and height prediction model

Bone age assessment is crucial in pediatric medicine to evaluate children’s skeletal maturity and growth, aiding in the early detection of developmental and endocrine disorders. Traditional methods are subjective and inconsistent. This study proposes a method using X-ray image processing and an enhanced Inception-V4 model, considering gender as a significant factor. We validated our model using large datasets and developed a height prediction model combining bone age prediction with patient attributes. Our model shows excellent performance in short and long-term height prediction. This research offers a more objective approach to assessing bone age, potentially benefiting children’s health by enabling earlier detection of developmental issues.