J. Wijckmans et al., "Parametric Modeling of the Human Body Using Kinect Measurements", in Proc. of 4th Int. Conf. on 3D Body Scanning Technologies, Long Beach CA, USA, 2013, pp. 117-126, http://dx.doi.org/10.15221/13.117.
Parametric Modeling of the Human Body Using Kinect Measurements
Jorn WIJCKMANS 1, Dorien VAN DEUN 1, Koen BUYS 2, Jos VANDER SLOTEN 1, Herman BRUYNINCKX 2
1 Dept. of Mechanical Engineering, Biomechanics Section, Katholieke Universiteit Leuven, Belgium; 2 Dept. of Mechanical Engineering, Division PMA, Katholieke Universiteit Leuven, Belgium
Personalized digital human modeling is useful for a wide variety of applications. An obvious interest comes from the entertainment industry, where movies and video games explore the possibilities of this technology. In biomechanics, human modeling assists in the design of person-specific solutions to improve the human well-being and ergonomics. Other applications exist in virtual dressing rooms, human-robot interactions in robotics, etc. Existing technology, although performing well, has the disadvantage of being expensive, immobile, not fully customizable and possibly requiring external body markers. To tackle these issues, this paper presents a modeling technique using the open source software MakeHuman, based on body measurements obtained with Microsoft's Kinect. The current solution is able to retrieve these measurements when the person is standing in a calibrated scene, this means when the person's position is known a priori. In order to retrieve the measurement data as a point cloud, and to process this point cloud, the PCL (Point Cloud Library) software is used, leading to a fully open source implementation. With these tools, solutions for person segmentation, measuring and personalized modeling are proposed. It appears that the current Kinect technology on itself is not very accurate for measuring body sizes. However, this work shows that the Kinect information combined with the MakeHuman modeling tool is valuable. The final model incorporates measures like body height, arm span, hip, waist and chest width, completed with information such as age, gender and weight. Evaluation of the resulting human model shows moderate to good results in modeling body height, hip and waist width, whereas chest width modeling is rather poor due to difficulties in chest width extraction from Kinect images.
Personalized digital human modeling, Kinect, MakeHuman, PCL
© Hometrica Consulting - Dr. Nicola D'Apuzzo, Switzerland, www.hometrica.ch.
Reproduction of the proceedings or any parts thereof (excluding short quotations for the use in the preparation of reviews and technical and scientific papers) may be made only after obtaining the specific approval of the publisher. The papers appearing in the proceedings reflect the author's opinions. Their inclusion in these publications does not necessary constitute endorsement by the editor or by the publisher. Authors retain all rights to individual papers.