智慧型家庭監控主要包含物件追蹤和事件偵測兩種技術。對物件偵測技術而言，particle filter和mean shift是最有名的兩個方法。Mean shift是效率較高的方法，但是有錯誤蔓延的問題。Particle filter可靠性較高，目標候選點愈多結果就會愈準確，但卻會增加時間複雜度。
在本論文中，我們提出結合物件切割的particle filter，其方法結合了以運動為基礎的物件切割法和particle filter的技術。以物件切割來說，估計總體運動參數可以分辨物體運動和攝影機運動，得到較粗略的物體遮罩。接下來，使用物件分類演算法自動分離個別的物件。藉由以高斯模型為主的切割結果，得到目標物大致的資訊，再利用particle filter與較少的目標候選點更新追蹤結果。對事件偵測的部份，我們針對跌倒事件，從追蹤結果中擷取特徵值。根據物體相對於攝影機的移動方向，使用不同的臨界值偵測跌倒事件。

An intelligent home surveillance mainly consists of object tracking and event detection. For object tracking, particle filter and mean shift are well-known methods. Mean shift is an efficient method, but Error propagation is a serious problem of mean shift. Particle filter is a robust tracking method and more target candidates would make the result more robust, but it would increase computation complexity.
In this thesis, we propose a segmentation-assisted particle filter for object tracking. Our method combines motion based segmentation and particle filter. For object segmentation, global motion parameters are estimated to distinguish local object motions from camera motions so as to obtain a rough object mask. Subsequently, an object clustering algorithm is used to automatically separate the individual video objects iteratively. Based on the rough segmentation result, we use particle filter with few target candidates to refine the tracking result. For event detection, we extract three features of each object for identifying and locating fall incidents. Our experiments show that the proposed method can correctly detect fall incidents in real time.

Chapter 1 3
Introduction 3
Chapter 2 8
Overview of Object Tracking 8
2.1 Object Representation 8
2.2 Similarity Measure 10
2.3 Particle Filter 10
2.4 Mean Shift 14
2.5 Hybrid Tracker Combining Mean Shift and Particle Filter 17
Chapter 3 19
Proposed Object Detection and Tracking 19
3.1 Initial Object Detection Using an Omni-Directional Camera 21
3.2 Segmentation-Assisted Particle Filter 25
A. Compressed-domain object segmentation 26
B. Label Set Method 29
C. Similarity Measure 31
D. Refinement of object tracking result using a particle filter 31
3.3 Experimental Results 33
Chapter 4 43
Fall Incident Detection for Intelligent Home Surveillance 43
4.1 Feature-Based Fall Incident Detection 44
4.1 Experimental Results 49
Chapter 5 54
Conclusions and Future Works 54
References 55

[1]C. S. Regazzoni, G. Fabri, and G. Vernazza, ed., Advanced Video-based Surveillance Systems, Kluwer Academic Publishers, 1999.
[2]G. L. Foresti,, P. Mahonen, and C. S. Regazzoni ed., Multimedia Video-based Surveillance Systems: Requirements, Issues and Solutions, Kluwer Academic Publishers, 2000.
[3]C. S. Regazzoni, V. Ramesh, and G. L. Foresti, “Scanning the issue/technology- Special issue on video communication, processing and understanding for third generation video surveillance systems,” Proc. IEEE, vol. 89, no. 10, pp. 1355-1367, Oct. 2001.
[4]W. Hu, T. Tan, L. Wang, and S. Maybank, “A survey on visual surveillance of object motion and behaviors,” IEEE Trans. Systems, Man, and Cybernetics- Part C: Applications and Reviews, vol. 34, no. 3, pp. 334-352, Aug. 2004.
[5]D. Comaniciu, V. Ramesh, and P. Meer, “Kernel-based object tracking,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 25, no. 5, pp. 564-577 May 2003.
[6]D. Comaniciu, V. Ramesh, and P. Meer, “Real-time tracking of non-rigid objects using mean shift,” in Proc. IEEE Int. Conf. Computer Vision & Pattern Recognition, vol.2, pp.142-149, June 2000
[7]D. Comaniciu and V. Ramesh, “Mean shift and optimal prediction for efficient object tracking,” in Proc. IEEE Int. Conf. Image Processing, vol. 3, pp. 70-73, Sept. 2000, Vancouver, Canada.
[8]K. Nummiaro, E. Koller-Meier, and L. Van Gool, “An adaptive color-based particle filter,” Image and Vision Computing, vol. 21, pp. 99-110, 2003.
[9]M. Isard and A. Blake, “Conditional density propagation for visual tracking,” Int. J. Computer Vision vol. 29, no. 1, pp. 5–28, 1998.
[10]C. Chang and R. Ansari, “Kernel particle filter for visual tracking,” IEEE Signal Processing Letters, vol. 12, no. 3, pp.242-245, Mar. 2005.
[11]E. Maggio and A. Cavallaro “Hybrid particle filter and mean shift tracker with adaptive transition model” in Proc. IEEE Int. Conf. Acoustics, Speech, and Signal Processing, pp. 3698-3701, Mar. 2005, Philadelphia, USA.
[12]K. Deguchi, O. Kawanaka, and T. Okatani “Object tracking by the mean shift of regional color distribution combined with the particle-filter algorithm,” in Proc. IEEE Int. Conf. Pattern Recognition, vol. 3, pp. 506-509, Aug. 2004, Cambridge, UK.
[13]C. Shan, Y. Wei, T. Tan and F. Ojardias, F “Real time hand tracking by combining particle filtering and mean shift,” in Proc. IEEE Int. Conf. Automatic Face and Gesture Recognition, pp. 669-674, May 2004, Seoul, Korea.
[14]T.-L. Liu and H.-T. Chen, “Real-time tracking using trust-region methods,” IEEE Trans. Pattern Anal. Machine Intell., vol. 26, no. 3, pp. 397-402, Mar. 2004.
[15]S.-C. Park, S.-H. Lim, B.-K. Sin, and S.-W. Lee “Tracking non-rigid objects using probabilistic Hausdorff distance matching,” Pattern Recognition, In press, available online April 2005.
[16]J. Gao, A. Kosaka, and A.-C. Kak “A multi-Kalman filtering approach for video tracking of human-delineated objects in cluttered environments,” Computer Vision and Image Understanding, vol. 99, no. 1, pp. 1-57, July 2005.
[17]D. Freedman and T. Zhang “Active contours for tracking distributions” IEEE Trans. Image Processing, vol. 13, no. 4, pp. 518-526, Apr. 2004.
[18]P. KaewTrakulPong and R. Bowden “A real time adaptive visual surveillance system for tracking low-resolution colour targets in dynamically changing scenes” Image and Vision Computing, vol. 21, pp.913-929, March 2003
[19]A. Cavallaro, O. Steiger, and T. Ebrahimi “Tracking video objects in cluttered background,” IEEE Trans. Circuits Syst. Video Technol., vol. 15, no. 4, Apr. 2005.
[20]T. Zhao and R. Nevatia “Tracking multiple humans in complex situations” IEEE Trans. Pattern Anal. Machine Intell., vol. 26, no. 9, pp.1208-1221, Sept. 2004.
[21]F. Porikli and O. Tuzel, Human body tracking by adaptive background models and mean-shift analysis, Technical Report, Mitsubishi Electric Research Laboratory, July 2003.
[22]T. Tamura et al., “An ambulatory fall monitor for the elderly,” in Proc. IEEE Int. Conf. EMBS, pp. 2608-2610, July 2000., Chicago, IL.
[23]G. Williams et al., “A smart fall & activity monitor for telecare applications,” in Proc. IEEE Int. Conf. EMBS, vol. 20, no. 3, pp. 1151-1154, 1998.
[24]N. Noury, T. Herve, V. Rialle, G. Virone, E. Mercier,G. Morey, A. Moro, and T. Porcheron, “Monitoring behavior in home using a smart fall sensor and position sensors,” in Proc. IEEE Int. Conf. Microtechnologies in Medicine and Biology, pp. 607-610, Oct. 2000, Lyon, France.
[25]N. Noury, “A smart sensor for the remote follow up of activity and fall detection of the elderly,” in Proc. IEEE Int. Conf. Microtechnologies in Medicine and Biology, pp. 314-317, May 2002, Lyon, France.
[26]C. Kim and J.-N. Hwang, “Object-based video abstraction for video surveillance systems,” IEEE Trans. Circuits Syst. Video Technol., vol. 12, no. 12, pp. 1128-1138, Dec. 2002.
[27]I. Haritaoglu, D. Harwood, and L. S. Davis, “W4: Real-time surveillance of people and their activities,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 22, no. 8, pp. 809-830, Aug. 2000.
[28]K. Yoon, D.F. Dementhon, and D. Doermann, “Event detection from MPEG video in the compressed domain,” in Proc. IEEE Int. Conf. Pattern Recognition, pp. 1819-1822, Barcelona, Spain, 2000.
[29]H. Nait-Charif and S. J. McKenna, “Activity summarisation and fall detection in a supportive home environment,” in Proc. IEEE Int. Conf. Pattern Rcognition, vol. 4, pp. 23-26, Aug. 2004, Cambridge UK.
[30]C.-W. Lin, Z.-H. Ling, Y.-C. Chang, and C.-J. Kuo, “Compressed-domain fall incident detection for intelligent home surveillance” in Proc. IEEE Int. Symp. Circuits Syst., pp. 3781-3784, May 2005, Kobe, Japan.
[31]D.J. Bullock and J.S. Zelek, “Real-time tracking for visual interface applications in cluttered and occluding situations” Image and Vision Computing, vol. 22, pp. 1083-1091, Mar. 2004
[32]C. Schell, S.P. Linder, and A.R. Zeidler, “Tracking highly maneuverable targets with unknown behavior” in Proc. IEEE, vol. 92, no. 3, pp. 558-574, March 2004
[33]Y. Su, M.-T. Sun, and V. Hsu, “Global motion estimation from coarsely sampled motion vector field and the applications,” in Proc. IEEE Int. Symp. Circuits Syst., vol. 2, pp. 628-631, Mar. 2003, Bangkok, Thailand.
[34]H. Zen, T. Hasegawa, and S. Ozawa, “Moving object detection from MPEG coded picture,” in Proc. IEEE Int. Conf. on Image Processing, Vol. IV, pp.25-29, Oct. 1999, Kobe, Japan.
[35]H.-L. Eng and K. –K. Ma, “Spatiotemporal segmentation of moving video objects over MPEG compressed domain,” in Proc. IEEE Int. Conf. Multimedia and Expo., vol.3, pp.1531 –1534, 2000, New York.
[36]M.-L. Jamrozik and M.-H. Hayes, “A compressed domain video object segmentation system,” in Proc. IEEE Int. Conf. Image Processing, vol.1, pp.113-116, Sep. 2002, New York, USA
[37]T. Kailath, “The divergence and Bhattacharyya distance measures in signal selection,” IEEE Trans. Comm. Technology, vol. 15, pp. 52-60, 1967.
[38]M. Wollborn and R. Mech, “Refined procedure for objective evaluation of video object generation algorithms,” Doc. ISO/IEC JTC1/SC29/WG11 M3448, Mar. 1998.