Bone Age Assessment Method Based on Deep Convolutional Neural Network

In this paper, the convolutional neural network based on deep learning theory is innovatively used to extract and classify the features of X-ray film of hand, and then to assess the age of bone. Based on the improved GoogleNet convolutional neural network, the method was applied to 301 cases of X-ray films of hand. To prevent the model from over-fitting, the training dataset is expanded 30 times by data augmentation method. After training, the model can achieve 91.3% of the bone age assessment accuracy on the test set. The classification performance validates the effectiveness of the proposed method. It provides theoretical basis and technical support for the research and implementation of a more accurate and rapid automated bone age assessment method, and provides a gauge and advice for physicians in the field of clinical, sports and judicial.