Real-Time Tracking of Articulated Human Models Using a 3D Shape-from-Silhouette Method
This paper describes a system, which acquires 3D data and tracks an eleven degree of freedom human model in real-time. Using four cameras we create a time-varying volumetric image (a visual hull) of anything moving in the space observed by all four cameras. The sensor is currently operating in a volume of approximately 500,000 voxels (1.5 inch cubes) at a rate of 25 Hz. The system is able to track the upper body dynamics of a human (x,y position of the body, a torso rotation, and four rotations per arm). Both data acquisition and tracking occur on one computer at a rate of 16 Hz. We also developed a calibration procedure, which allows the system to be moved and be recalibrated quickly. Furthermore we display in real-time, either the data overlaid with the joint locations or a human avatar. Lastly our system has been implemented to perform crane gesture recognition.
KeywordsJoint Angle Visual Hull Joint Location Eleven Degree Human Avatar
Unable to display preview. Download preview PDF.
- 1.C. Bregler, J. Malik: Tracking people with twists and exponential maps. Proc. IEEE CVPR, (1998).Google Scholar
- 3.C. Wren, A. Azarbayejani, T. Darrel, A. Pentland: Pfinder: real-time tracking of the human body. IEEE Transactions PAMI, 19 (1997) 780–785.Google Scholar
- 4.I. A. Kakadiaris, D. Metaxas: 3D human body model acquisition from multiple views. Proc. 5th Int. Conf. on Computer Vision, Boston, MA (June 1995) 618–623.Google Scholar
- 5.D. Gravrila, L. Davis: 3-D model-based tracking of humans in action: a multi-view approach. In: Proc. IEEE CVPR, San Francisco (1996).Google Scholar
- 6.K. M. Cheung, T. Kanade, J. Y. Bouguet, M. Holler: A real time system for robust 3D voxel reconstruction of human motions. In: Proc. IEEE CVPR, (2000).Google Scholar
- 7.D. Snow, P. Viola, R. Zabih: Exact voxel occupancy with graph cuts. In: Proc. IEEE CVPR, (2000).Google Scholar