Upper Body Tracking for Interactive Applications
In this paper we describe a complete method for building a perceptual user interface in indoor uncontrolled environments. The overall system uses two calibrated cameras and does initialization: it detects user, takes his/her measurements, builds a 3D-Model. It performs matching/tracking for: trunk, head, left arm, right arm and hands. The system is waiting for a user in a predefined posture, once the user has been detected he/she is analysed to take measurements are taken and a 3D-Model is built. Tracking is carried out by a Particle Filter algorithm splited in three steps: tracking of head-trunk, tracking of left arm and tracking of right arm. This proposed divide and conquer solution improves computation time without getting better or similar results than sequential solution. The matching process uses two sub-matching functions, one to compute color and another to compute shape one. Finally the system provides numerical values for joints and end effectors to be used for interactive applications.
KeywordsParticle Filter Match Function Skin Region Tracking Process Interactive Application
Unable to display preview. Download preview PDF.
- 1.Krahnstoever, N., Yeasin, M., Sharme, R.: Automatic acquisition and initialization of articulated models. Machine Vision and Applications 14(4) (2003)Google Scholar
- 2.Allen, B., Curless, B., Popovic, Z.: The space of human body shapes: reconstruction and parameterization from range images. In: Proc. ACM SIGGRAPH, pp. 587–594 (2003)Google Scholar
- 3.Carranza, J., Theobalt, C., Magnor, M., Seide, H.P.: Free-viewpoint of human actors. In: Proc. ACM SIGRRAPH, pp. 565–577 (2003)Google Scholar
- 5.Deustcher, J., Blake, A., Reid, I.: Articulated body motion capture by annealed particle filtering. In: IEEE Conf. Computer Vision and Pattern Recognition, vol. 2, pp. 126–133 (2000)Google Scholar
- 6.Mitchelson, J., Hilton, A.: Hierachical tracking of multiple people. In: British Machine Vision Conference (2003)Google Scholar
- 7.Buades, J.M., Gonzalez, M., Perales, F.J.: A New Method for Detection and Initial Pose Estimation based on Mumford-Shah Segmentation Functional. In: Perales, F.J., Campilho, A.C., Pérez, N., Sanfeliu, A. (eds.) IbPRIA 2003. LNCS, vol. 2652, pp. 117–125. Springer, Heidelberg (2003)CrossRefGoogle Scholar