Bone Age Classification Using Deep Learning: A ResNet50 Model Approach

The aim of this study was to create and assess a deep learning based automated Bone Age detection model using ResNet50 architecture. Bone age assessment is critical to diagnosing and treating pediatric conditions so our goal was to create a model that could accurately predict bone age from medical images. A transfer learning approach was used to use the pre trained ResNet50 model with a specific task of bone age classification, which was fine tuned. The dataset that was used was 12,611 medical images and were grouped into 8 bone age groups. A confusion matrix was used to show accuracy, precision, recall, F1 score, etc. to evaluate the performance on the basis of different age groups and the model was trained and evaluated using standard classification metrics. On average, the model had $\mathrm{9 3. 1 2 \%}$ overall accuracy with very high precision, recall and F1 scores on most classes. The few misclassifications in the confusion matrix demonstrated strong generalization across age groups. This work shows that deep learning models, like ResNet50, can automate bone age assessment. The model performs exceptionally well, potentially making it suitable for use in clinical settings where it could help clinicians make faster, more accurate diagnoses. The results emphasise the utility of applying artificial intelligence in medical imaging for the development of future, more sophisticated, automated diagnostic techniques.