Estimation of Femur Measurement of Malaysian Adults Using the Artificial Neural Network

Artificial Neural Network (ANN) method is used to estimate femur bone for Malaysian adults’ population. The main objective of this study is to investigate the reliability of ANN to predict the length of femur for Malaysia adults. Currently computerized tomography (CT) scan and Magnetic Resonance Imaging (MRI) method were used to obtain multilayer images which the images converted to 3D image of bones for the measurements purposes. These methods are not safe which may harm hearing, claustrophobia and anxiety, peripheral muscle, and nerve stimulation due to the amount of radiation exposure during CT Scan or MRI. In addition, CT scans usually require more exposure to radiation than common x-rays because they use a series of x-ray images. Increased exposure means a slightly higher risk of possible short-term and long-term health effects. Therefore, as alternative method, ANN is chosen which as far as we concern, ANN is just a software based. However, to obtain the reliable ANN model, the measurements data are needed to train, validate and testing it. Thus, a total of 100 femur bones for normal Malaysian adults is taken from CT scan data to train, validate and test the ANN model. Based on the performance result, the ANN is capable of predict the measurement of femur bone with a high precision and accuracy since the percentage of errors are below than 5%. The purpose of this study holds the potential to serve as a valuable asset for the prediction of surgical outcomes and the analysis of risk factors for femur bone repair in Malaysia with minimal adverse effect to human body during bone measurements session.