在本篇論文中，我們發展了一個在兩正交視野下以模型為基礎的即時人體運動參數分析系統。我們使用兩台攝影機去抓取正面和側面的人體運動資訊，為了追蹤人體的動作，我們對雙手和雙腿使用了兩種不同的方法去估算人體運動時的關節角度，首先，我們使用Kalman Filter去預測並修正雙手的姿勢，然後再透過比較影像中前景物體與二維人體模型的相似程度，調整出最佳的人體腿部關節角度，然後再整合雙手和雙腿的關節角度來得到一個完整的動作。我們的人體運動分析系統分成巨觀運動分析和微觀運動分析，巨觀運動分析經由分析二元影像的垂直投影和側投影得到雙手和雙腿動作的基本資訊，然後再使用Kalman Filter做微觀運動分析去追蹤雙手動作的關節角度，以及用模型比對的方式去找到雙腿動作的關節角度。

In this thesis, we introduce a real time two orthogonal views human motion analysis system. We use two cameras to capture the facade and flank views of the human motion. To track the motion of the human object, we propose two methods to estimate the motion parameters (BAPs) of the human arm and leg. First, we use Kalman filter to predict arms posture. Second, we track the leg by model-based matching method. The human motion analysis is divided into macro motion analysis and micro motion analysis. The former identifies the certain well-defined postures and the latter traces the variation of joint angle parameters. Before tracking, we analyze the vertical projection profile and horizontal projection profile in each view to identify postures. With the identified postures, we can apply the Kalman filter to track the motion of joint angels and generate the BAPs.

[1] G. Johansson, “Visual motion perception,” Sci. Am. 232(6), 1975, 76-88.
[2] C. Y. Chuang and C. L. Huang, “A real time model-based human motion analysis system in multiple-view,” Master thesis, Department of Electrical Engineering in NTHU, Taiwan, 2001.
[3] T. H. Tzeng and C. L. Huang, “A model-based human motion analysis system in multiple-view,” Master thesis, Department of Electrical Engineering in NTHU, Taiwan, 2003.
[4] Osama Masoud, Nikos Papanikolopoulos, “A method for human action recognition,” Image and Vision Computing 21 (2003), 729-743
[5] S. L. Dockstader and A. M. Tekalp, “On the tracking of articulated and occluded video object motion,” Real-Time Imaging 7, 415-432 (2001).
[6] S. L. Dockstader and A. M. Tekalp, “Multiple camera tracking of interacting and occluded human motion,” Proc. IEEE, vol. 89, no. 10, pp.1441-1455, Oct. 2001.
[7] S. L. Dockstader, M. J. Berg and A. M. Tekalp, “Stochastic kinematic modeling and feature extraction for gait analysis,” IEEE TRANSACTIONS ON IMAGE PROCESSING, vol. 12, no. 8, August 2003.
[8] R. E. Kalman, “A new approach to linear filtering and prediction problems,” Transaction of the ASME—Journal of Basic Engineering, pp.35-45, March. 1960.
[9] T. J. Broida, S. Chandrashekhar, and R. Chellappa, “Recursive 3-D motion estimation from a monocular image sequence,” IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS, vol. 26, no. 4, July. 1990.
[10] Ali Azarbayejani, Bradley Horowitz, and Alex Pentland, Perceptual Computing Section, MIT Media Laboratory, 1993.
[11] Simon Haykin, “Kalman filtering and neural networks,” John Wiley&Sons, INC., New York, 2001
[12] Berthold K. P. Horn, “Closed-form solution of absolute orientation using unit quaternions,” Reprinted from Journal of the Optical Society of America A, vol. 4, page 629, April 1987
[13] N. Krahnstover, M. Yeasin, and R. Sharma, “Toward a unified framework for tracking and analysis of human motion,” IEEE Workshop on Detection and Recognition Events in Video, 2001
[14] Thanart Horprasert, David Harwood, and Larry S. Davis, “A robust background subtraction and shadow detection,” Proceedings of the Fourth Asian Conference on Computer Vision, page 983-988, 2000
[15] I. C. Chang and C. L. Huang, “The model-based human body motion analysis system,” Image and Vision Computing, vol.18, pp.1067-1083, 2000.
[16] Greg Welch and Gary Bishop, “An introduction to the Kalman filter,” Department of Computer Science University of North Carolina at Chapel Hill Chapel Hill, NC 27599-3175, Friday, May 23, 2003
[17] Mobinder S.Grewal and Angus P.Andrews, “Kalman Filtering Theory and Practice Using MATLAB,” John Wiley&Sons, INC.