Isam F AbuQasmieh and Hiam H. Al-Quran
body weight, body height, torso area, convex hull
Weight can provide useful information for identification and medical purposes, such as in drug delivery control and anesthetic dose calculations. In this study, the weight estimation was obtained from the measurements and proportions of the human body images. Data are extracted from the segmentation and analysis of the binary body’s image parts for the subject under study. The experimental work has been performed on healthy adult subjects under ideal conditions. 85% of the database has been exploited for training and building the 2D (surface) polynomial model using surface fitting. The rest 15% has been used for investigating and testing the validity of the built models. For building multiple models of surface polynomial, surface fitting was performed between the measured weights of the training set as the dependent variable and two of the four anthropometric data as independent variables. The anthropometric features were extracted from the segmented body regions using different image processing techniques with novel corners detection technique. The performance analysis of our approach shows that the proposed system provides an accurate automatic estimation of the human body weight.