Improving Accuracy for Bone Age Prediction from X-Ray Image using Convolutional Neural Network Technique over K-Nearest Neighbors

Abstract

Aim: To enhance accuracy in predicting bone age from x-ray image to that of Chronological ages using Novel Convolutional Neural Network technique in comparison with K-Nearest Neighbor. Materials and methods: Classification is performed by a convolutional neural network (N=10) Over a Support Vector Machine (N=10). The sample size is calculated using Gpower with pretest power 0.8 as an alpha 0.005 Result: Mean accuracy Convolutional Neural Network (82.36%) is high compared to K-Nearest Neighbor (70.26%). The significance value for accuracy and loss is 0.389 (p>0.05) Conclusion: The mean accuracy of the college recommendation system in Convolutional Neural Networks is better than the K-Nearest Neighbor.