We used open source tools to segment each body scan into 16 segments (head, torso, abdomen, pelvis, left and right hand, forearm, upper arm, foot, shank and thigh) and wrote custom software to estimate each segment’s mass, mass moment of inertia in the three principal orthogonal axes relevant to the center of the segment, longitudinal length, and center of mass. To evaluate our system, we performed repeated 3D scans of 21 healthy participants (10 male, 11 female). Using a Microsoft Kinect Version 2 camera, we developed a 3D surface scanning protocol that enabled the estimation of participant-specific body segment parameters. Our objective was to develop a participant-specific 3D scanning and body segmentation method that estimates body segment parameters without any assumptions about the geometry of the body, ethnic background, and gender, is low-cost, fast, and can be readily available. Commonly used methods for estimating participant-specific body segment parameters are either expensive and out of reach (medical imaging), have many underlying assumptions (geometrical modelling) or are based on a specific subset of a population (regression models). Participant-specific estimates of body segment parameters are desirable as this requires fewer prior assumptions and can reduce outcome measurement errors.
Body segment parameters are inputs for a range of applications.
