Automated Calculation of Height and Area of Human Body for Estimating Body Weight Using a Matlab-based Kinect Camera

Abstract

ABSTRACT Body size is an important parameter in radiology examinations. Measurements of body size may be carried out remotely using a Kinect camera. We aim to develop a Matlab-based software that can be used to automatically measure the height and area in order to estimate the weight of a human body without direct contact. The size and depth of the image were calibrated for various object-Kinect distances. The depth image was segmented and normalized into a binary image, from which the height and area were calculated. The height and area of the body were correlated with the body weight. We analyzed the images obtained from 147 participants aged 5–70 years old and weighing 13.7–89.9 kg. The height measured using the Kinect camera had a high correlation with a ruler with R2 = 0.96 (p = 0.0096), and error percentage of 1.04%. It is also found that the height and area of the participants are well correlated with the weight of the participants with R2 of 0.453 and 0.658, respectively. A graphical user interface for automatically calculating the height and area of the human body has been successfully developed.