Particle Swarm Optimization Based Support Vector Machine for Human Tracking
Human tracking is one of the most important researches in computer vision. It is quite useful for many applications, such as surveillance systems and smart vehicle systems. It is also an important basic step for content analysis for behavior recognition and target detection. Due to the variations in human positions, complicated backgrounds and environmental conditions, human tracking remains challenging work. In particular, difficulties caused by environment and background such as occlusion and noises should be solved. Also, real-time human tracking now seems a critical step in intelligent video surveillance systems because of its huge computational workload. In this paper we propose a Particle Swarm Optimization based Support Vector Machine (PSO-SVM) to overcome these problems. First, we finish the preliminary human tracking step in several frames based on some filters such as particle filter and kalman filter. Second, for each newly come frame need to be processed, we use the proposed PSO-SVM to process the previous frames as a regression frame work, based on this regression frame work, an estimated location of the target will be calculated out. Third, we process the newly come frame based on the particle filter and calculate out the target location as the basic target location. Finally, based on comparison analysis between basic target location and estimated target location, we can get the tracked target location. Experiment results on several videos will show the effectiveness and robustness of the proposed method.
KeywordsHuman tracking Occlusion Real-time Particle filter PSO-SVM
This work was supported in part by Nanjing Audit University, Nanjing, Jiangsu, China and “six kinds of peak talents” high level support of Jiangsu.
- 2.Isard, M., Blake, A.: Contour tracking by stochastic propagation of conditional density. Computer Vision-ECCV’96, Lecture Notes in Computer Science, vol. 1064/1996, pp. 343–356 (1996)Google Scholar
- 4.Zhai, Y., Yeary, M.B., Cheng, S., Kehtarnavaz, N.: An object-tracing algorithm based on multiple-model particle filtering with state partitioning. IEEEE Trans. Instrum. Meas. 58, 1797–1809 (2009)Google Scholar
- 5.MacCormick, J., Blake, A.: A probabilistic exclusion principle for tracking multiple objects. In: Proceedings of the International Conference on Computer Vision, pp. 572–578 (1999)Google Scholar
- 7.Ye, L., Zhang, Q., Guan, L.: Use hierarchical genetic particle filter to figure articulated human tracking. In: IEEE International Conference on Multimedia and Expo, pp. 1561–1564 (2008)Google Scholar
- 10.Ristic, B., Arulampalam, S., Gordon, N.: Beyond the Kalman Filter: Particle Filters for Tracking Applications. Artech House (2004)Google Scholar
- 12.Smola, A., Vapnik, V.: Support vector regression machines. Advances in Neural Information Processing Systems, vol. 9, pp. 155–161. MIT Press (1997)Google Scholar
- 14.Wang, J., Xing, Y., Cheng, L., Qin, F.: The Prediction of Mechanical Properties of Cement Soil Based on PSO-SVM, Computational Intelligence and Software Engineering (CiSE), 2010 International Conference on, pp. 10–12 (2010). doi: 10.1109/CISE.2010.5677256