In this research, we designed a human tracking system. In the system, firstly, we have developed a spatially-extended background model for foreground detection. In the background model, we have used the probabilities of joint random vectors between near pixels to model the spatial relations. To reduce the cost of modeling the pixel-pairs, we calculate the mutual information in each pixel-pair for finding the spatially-dependent pixel-pairs.
In general environments, when the background regions are stable, the Gaussian background model is suitable to segment foreground regions. However, when background regions change, the model is unsuitable. To detect foreground regions more accurately with respect to either changed or still background regions, we should combine our propose model with Gaussian background model. To achieve this, some heuristic rules should be created for deciding which model should be selected. This is left for future studies.
To track a human, we decompose a human body into three parts, the head, torso, and hip-leg, and use color-based particle filters to track the three parts separately. We combined the appearance models of target object and background scene to calculate the weight of each particle. To reduce the redundancy during calculating color histograms in the overlapped regions of the particles, we have created a cumulative histogram map for each frame. We have also proposed an SVM-based method to detect the lost tracking part. In the tracking algorithm, we have used a particle filter for tracking an individual part. Since the particle number affects the tracking performance and tracking speed, we use entropy of particle weights to modify the
76
particle number dynamically. To further improve the tracking accuracy, we have designed a histogram equalization method for color histogram comparison. The experimental results show that the three parts tracking algorithm can improve the tracking accuracy significantly.
In this research, we assume that a human is standing up and the three body parts can be segmented from top to bottom. If a human crouches down or lies down, the body part decomposition may fail. Our experimental results show that in the case of failure, the three parts will not be labeled correctly. However, the failure will be adjusted after the target person stands up again, since we can detect the failure by SVM. To improve the body part decomposition, we may train detectors for different body parts. This is left for future research.
In our method, each of the three parts is tracked by a particle filter independently.
The relative positions of the body parts are used to detect the tracking failure. We can reduce tracking failures by preventing the particles of abnormal poses to be generated.
To achieve the goal, we needs to combine the state vectors of the three parts into a single vector to be tracked by a particle filter. Then the particle weights are adjusted according to the relative positions of the body parts. Also the behaviors of intruders defined on object appearances and the trajectories found will be analyzed. These are all left for future research.
77
Bibliography
[1] L. Wang, W. Hu, and T. Tan, "Recent developments in human motion analysis,"
Journal of Pattern Recognition, vol. 36, no. 3, pp. 585-601, Mar. 2003.
[2] I. Haritaoglu, D. Harwood, and L. S. Davis, "W4: Real-Time Surveillance of People and Their Activities," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, no. 8, pp. 809-830, Aug. 2000.
[3] C. R. Wren, A. Azarbayejani, T. Darrell, and A. P. Pentland, "Pfinder: Real-Time Tracking of the Human Body," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 19, no. 7, pp. 780-785, Jul. 1997.
[4] R. Cucchiara, C. Grana, M. Piccardi, and A. Prati, "Detecting moving objects, ghosts, and shadows in video streams," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 10, pp. 1337-1342, Oct. 2003.
[5] Y. N. T. T. M. H. Shiry Ghidary, "Human detection and localization at indoor environment by home robot," in Proceeding of IEEE International Conference on Systems, Man, and Cybernetics, vol. 2, Nashville, USA, 2000, pp. 1360-1365.
[6] A. Elgammal, R. Duraiswam, D. Harwood, and L. Davis, "Background and Foreground Modeling Using Nonparametric Kernel Density Estimation for Visual Surveillance," Proceedings of the IEEE, vol. 90, no. 7, pp. 1151-1163, Jul.
2002.
78
[7] S. J. McKenna, S. Jabri, Z. Duric, A. Rosenfeld, and H. Wechsler, "Tracking groups of people," Computer Vision and Image Understanding, vol. 80, no. 1, pp.
42-56, Oct. 2000.
[8] A. Prati, I. Mikic, M. M. Trivedi, and R. Cucchiara, "Detecting moving shadows:
algorithms and evaluation," IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 25, no. 7, pp. 918-923, Jul. 2003.
[9] C. Stauffer and W. E. L. Grimson, "Learning patterns of activity using real-time tracking," IEEE Transactions on Pattern Analysis and Machine Intelligence , vol.
22, no. 8, pp. 747-757, Aug. 2000.
[10] D.-S. Lee, "Effective Gaussian mixture learning for video background subtraction," IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 27, no. 5, pp. 827-832, May 2005.
[11] H. Wang and D. Suter, "A re-evaluation of mixture-of-Gaussian background modeling," in Proceeding of IEEE International Conference on Acoustics, Speech, and Signal Processing, vol. 2, PA, USA, 2005, pp. 1017-1020.
[12] K. Kim, T. H. Chalidabhongse, D. Harwood, and L. Davis, "Real-time foreground-background segmentation using codebook model," Real-Time Imaging, vol. 11, no. 3, pp. 172-185, Jun. 2005.
[13] D. R. Magee, "Tracking multiple vehicles using foreground, background and motion models," Image and Vision Computing, vol. 22, no. 2, pp. 143-155, Feb.
2004.
[14] E. Durucan and T. Ebrahimi, "Change detection and background extraction by linear algebra," Proceedings of the IEEE , vol. 89, no. 10, pp. 1368-1381, Oct.
2001.
79
[15] L. Li, W. Huang, I. Y.-H. Gu, and Q. Tian, "Statistical modeling of complex backgrounds for foreground object detection," IEEE transactions on image processing , vol. 13, no. 11, pp. 1459-1472, Nov. 2004.
[16] L. Li and M. K. H. Leung, "Integrating intensity and texture differences for robust change detection," IEEE Transactions on Image Processing, vol. 11, no. 2, pp. 105-112, Feb. 2002.
[17] Y. Wang, T. Tan, K.-F. Loe, and J.-K. Wu, "A probabilistic approach for foreground and shadow segmentation in monocular image sequences," Pattern Recognition, vol. 38, no. 11, pp. 1937-1946, Nov. 2005.
[18] P. Pérez, C. Hue, J. Vermaak, and M. Gangnet, "Color-Based Probabilistic Tracking," in Proceeding of 7th European Conference on Computer Vision, vol.
1, Copenhagen, Denmark, 2002, pp. 661-675.
[19] K. Nummiaro, E. Koller-Meier, and L. V. Gool, "An adaptive color-based particle filter," Image and Vision Computing, vol. 21, no. 1, pp. 99-110, Jan. 2003.
[20] P. Viola and M. Jones, "Rapid Object Detection using a Boosted Cascade of Simple Features," in Proceeding of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 1, Los Alamitos, CA, USA, 2001, pp. 511-518.
[21] A. Mohan, C. Papageorgiou, and T. Poggio, "Example-based object detection in images by components," IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 23, no. 4, pp. 349-361, Apr. 2001.
[22] C. J. C. Burges, "A Tutorial on Support Vector Machines for Pattern Recognition," Data Mining and Knowledge Discovery, vol. 2, no. 2, pp. 121-167, Jun. 1998.
80
[23] H. Deng and D. A. Clausi, "Unsupervised image segmentation using a simple MRF model with a new implementations scheme," Pattern Recognition, vol. 37, no. 12, pp. 2323-2335, Dec. 2004.
[24] T. B. Moeslund, A. Hilton, and V. Krüger, "A survey of advances in vision-based human motion capture and analysis," Computer Vision and Image Understanding, vol. 104, no. 2-3, pp. 90-126, Nov. 2006.
[25] W. Hu, T. Tan, L. Wang, and S. Maybank, "A survey on visual surveillance of object motion and behaviors," IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews, vol. 34, no. 3, pp. 334-352, Aug.
2004.
[26] A. Yilmaz, O. Javed, and M. Shah, "Object tracking: A survey," ACM Computing Surveys, vol. 38, no. 4, pp. 1-45, Dec. 2006.
[27] T. B. Moeslund and E. Granum, "A Survey of Computer Vision-Based Human Motion Capture," Computer Vision and Image Understanding, vol. 81, no. 3, pp.
231-268, Mar. 2001.
[28] D. M. Gavrila, "Visual analysis of human movement: A survey," Computer Vision and Image Understanding, vol. 73, no. 1, pp. 82-98, Jan. 1999.
[29] H. Sidenbladh, "Detecting human motion with support vector machines," in Proceeding of International Conference on Pattern Recognition, vol. 2, Cambridge, UK, 2004, pp. 188-191.
[30] Y. Ivanov, A. Bobick, and J. Liu, "Fast Lighting Independent Background Subtraction," International Journal of Computer Vision, vol. 37, no. 2, pp.
199-207, Jun. 2000.
[31] A. S. Micilotta, E. J. Ong, and R. Bowden, "Detection and tracking of humans by
81
probabilistic body part assembly," in Proceeding of British Machine Vision Conference, Oxford, UK, 2005, pp. 429-438.
[32] D. Comaniciu, V. Ramesh, and P. Meer, "Kernel-Based Object Tracking," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 5, pp.
564-575, May 2003.
[33] L. Zhao and C. Thorpe, "Stereo-and neural network-based pedestrian detection,"
IEEE Transactions on Intelligent Transportation Systems, vol. 1, no. 3, pp.
148-154, Sep. 2000.
[34] C. Shan, T. Tan, and Y. Wei, "Real-time hand tracking using a mean shift embedded particle filter," Pattern Recognition, vol. 40, no. 7, pp. 1958-1970, Jul.
2007.
[35] E. Polat, M. Yeasin, and R. Sharma, "Robust Tracking of Human Body Parts for Collaborative Human Computer Interaction," Computer Vision and Image Understanding, vol. 89, no. 1, pp. 44-69, Jan. 2003.
[36] S. L. Dockstader and N. S. Imennov, "Prediction for human motion tracking failures," IEEE Transactions on Image Processing, vol. 15, no. 2, pp. 411-421, Feb. 2006.
[37] Q. Zhou and J. K. Aggarwal, "Object tracking in an outdoor environment using fusion of features and cameras," Image and Vision Computing, vol. 24, no. 11, pp. 1244-1255, Nov. 2006.
[38] M. Isard and A. Blake, "CONDENSATION - Conditional Density Propagation for Visual Tracking," International Journal of Computer Vision, vol. 29, no. 1, pp. 5-28, Aug. 1998.
[39] M. S. Arulampalam, S. Maskell, N. Gordon, and T. Clapp, "A Tutorial on Particle
82
Filters for Online Nonlinear/Non-Gaussian Bayesian Trackin," IEEE Transactions on Signal Processing, vol. 50, no. 2, pp. 174-188, Feb. 2002.
[40] D. A. Forsyth and M. M. Fleck, "Body Plans," in Proceeding of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Puerto Rico, 1997, pp. 678--683.
[41] A. Shashua, Y. Gdalyahu, and G. Hayun, "Pedestrian detection for driving assistance systems: single-frame classification and system level," in Proceeding of IEEE Intelligent Vehicles Symposium, Parma, Italy, 2004, pp. 1-6.
[42] "Human Detection Based on a Probabilistic Assembly of Robust Part Detectors,"
in Proceeding of European Conference on Computer Vision, Prague, Czech Republic, 2004, pp. 69-82.
[43] S. Ioffe and D. Forsyth, "Probabilistic Methods for Finding People,"
International Journal of Computer Vision, vol. 43, no. 1, pp. 45-68, Jun. 2001.
[44] C. Chang, R. Ansari, and A. Khokhar, "Efficient tracking of cyclic human motion by component motion," IEEE Signal Processing Letters, vol. 11, no. 12, pp.
941-944, Dec. 2004.
[45] T. J. Roberts, S. J. McKenna, and I. W. Ricketts, "Human pose estimation using learnt probabilistic region similarities and partial configurations," in Proceeding of European Conference on Computer Vision, Prague, Czech Republic, 2004, pp.
291-303.
[46] D. Ramanan, D. A. Forsyth, and A. Zisserman, "Strike a Pose: Tracking People by Finding Stylized Poses," in Proceeding of Computer Vision and Pattern Recognition, vol. 1, Washington, DC, USA, 2005, pp. 271-278.
[47] B. Wu and R. Nevatia, "Detection and Tracking of Multiple, Partially Occluded
83
Humans by Bayesian Combination of Edgelet based Part Detectors,"
International Journal of Computer Vision, vol. 75, no. 2, pp. 247-266, Nov. 2007.
[48] C. Lerdsudwichai, M. Abdel-Mottaleb, and A.-N. Ansari, "Tracking Multiple People with Recovery from Partial and Total Occlusion," Pattern Recognition, vol. 38, no. 7, pp. 1059-1070, Jul. 2005.
[49] S. Khan and M. Shah, "Tracking people in presence of occlusion," in Proceeding of Asian Conference on Computer Vision, Taipei, Taiwan, 2000, pp. 1132-1137.
[50] C. Chow and C. Liu, "Approximating discrete probability distributions with dependence trees," IEEE Transactions on Information Theory, vol. 14, no. 3, pp.
462-467, May 1968.
[51] D. Fox, "Adapting the Sample Size in Particle Filters Through KLD-Sampling,"
The International Journal of Robotics Research, vol. 22, no. 12, pp. 985-1003, 2003.
[52] CAVIAR. [Online]. http://www.dai.ed.ac.uk/homes/rbf/CAVIAR