Early Prediction of Osteoporosis using Signal Processing Technique

Abstract

Osteoporosis is a very common not so known disease which affects millions of people throughout the year. With the advancement in the Medical Field, it is desired to use machine learning algorithms to get some breakthrough in the detection of this bone disease. Therefore, the aim of this research work is to predict the osteoporosis from the clinical images using a machine learning algorithm. The signal processing technique used here is Machine Learning. A Simple Convolutional Neural Network (CNN) is used for this study. A dataset of 953 normal images and 881 osteoporosis images are used. All clinical images used are X-Ray images. Then image classification is used to classify the image and to predict the correct output. Osteoporosis accuracy is an important metric used here for prediction. The proposed CNN model achieved the training accuracy up to 93% and validation accuracy of up to 91% and helps in prediction of a category of image.