Deep Learning for Osteoporosis Classification using Convolutional Neural Network on Knee Radiograph Dataset to Compare Classification Accuracy between RGB Images and Grayscale Images
Author(s): Usman Bello Abubakar, Moussa Mahamat Boukar, Steve Adeshina and Senol Dane*
Abstract
Osteoporosis is a silent killer disease in the elderly that causes bone fragility and fractures. An early and precise diagnosis of osteoporosis is necessary to save the patient's life. Osteoporosis, along with bone fracture: its clinical manifestation is a complex illness towards which a great deal of research has been made upon. Machine Learning (ML) and Deep Learning (DL) advancements have allowed the area of Artificial Intelligence (AI) to make significant progress in complicated data environments where humans' capacity to find high-dimensional connections is limited. This work focuses on determining the appropriate image format to be used by a deep learning model to predict osteoporosis from knee radiograph. The image format considered in the study is RGB images and Grayscale images. Two Convolutional Neural Network (convent) of the same structure were trained on set of same images with RGB and grayscale. The result showed that the convent trained on grayscale images had a better accuracy than the convent trained on RGB images.