Articulated Human Motion Tracking by Sequential Annealed Particle Swarm Optimization
In this paper, we present a novel generative method for articulated human motion tracking. The principle contribution is the development of a modified Particle Swarm Optimization (PSO) algorithm for pose optimization in latent space of human motion. There are three characteristics in the proposed method. Firstly, we learn the latent space of human motion using PCA and perform human motion analysis in this latent space, which results to be more efficient and accurate. Secondly, we introduce simulated annealing into traditional PSO. A new algorithm, termed annealed PSO (APSO) is designed for pose optimization, which can get global optimum solution more efficiently. Lastly, we apply APSO for human pose estimation. And a sequential APSO (SAPSO) method is proposed for motion tracking. Experimental results on different motion types and different image sequences show that our method achieves better results than state-of-art methods.
KeywordsPose estimation Motion tracking Particle swarm optimization
Unable to display preview. Download preview PDF.
- 1.Sminchisescu, C.: Human motion understanding, modeling, capture and animation. In: Kleete, R., Metaxas, D., Rosenhahn, B. (eds.) 3D Human Motion Analysis in Monocular Video, Techniques and Challenges. Springer (2007)Google Scholar
- 6.Howe, N.R., Leventon, M.E., Freeman, W.T.: Bayesian Reconstruction of 3D human motion from single-camera video. In: Advances in Neural Information Processing Systems, pp. 820–826. IEEE Press, Denver (2000)Google Scholar
- 9.Zhang, X.Q., Hu, W.M., Maybank, S., Xi, L.: Sequential particle swarm optimization for visual tracking. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 23–28. IEEE Press, Anchorage (2008)Google Scholar
- 10.CMU Motion Capture database, http://mocap.cs.cmu.edu/
- 11.Ormoneit, D., Sidenbladh, H., Black, M.J., Hastie, T.: Learning and tracking cyclic human motion. In: Advances in Neural Information Processing Systems, pp. 894–900. IEEE Press, Vancouver (2001)Google Scholar