Automated Hand X-Ray Based Gender Classification and Bone Age Assessment Using Convolutional Neural Network

Abstract

Bone Age Assessment (BAA) is a medical approach to predict the growth in any individual and for this gender the classification has immense importance in medical research and forensics. To the best of our knowledge we have introduced a novel framework, which classifies the gender and predict the age of that individual by using a single left-hand radiograph. Deep Convolutional Neural Network (DCNN) as a method of learning and predicting the results gave us the accuracy of 79.6% for gender classification and for age classification we have achieved MAD 0.50 years and RMS 0.67 years. We have studied the methods of transfer learning and trained our dataset with VGG-16 model to find the optimal solution.