異常行為事件之偵測在視覺監視系統中係一項重要的研究課題，尤其在擁擠環境之影像中，準確且可信賴的追蹤和偵測往往是一項重大挑戰。本論文將考量連續影像中時間和空間的特徵關係，基於社會力模型(Social Force Model)的概念而提出HOSF方法作為特徵描述，來解決傳統基於物件路徑之偵測方法不適用於擁擠影像的問題。經由粒子的分割建立cuboids，再透過HOSF特徵向量訓練出判斷事件異常性所依據的字典(dictionary)，最後根據z-value計算事件與字典中codeword的相似情形來決定影像中事件的異常與否。
本論文所提方法包含以下特點：（1）訓練過程為全自動，不需事先經由人工決定事件屬於正常或異常；（2）特徵描述採用粒子和社會力，不需事先對物件做追蹤，因此不論擁擠人群或人少的場景皆可適用；（3）使用z-value方法來估算事件的異常性，由於計算簡單，訓練完成後即可達到即時偵測。

In this paper a simple and effective crowd behavior normality method is proposed. Feature vector, so called HOSF (histogram of oriented social force), and consists of concatenating local histogram of oriented social force. A dictionary of codewords is trained to include typical HOSF. To detect whether an event is normal is accomplished by comparing how similar to the closest codeword via z-value. The proposed method includes the following characteristic: (1) the training is automatic instead of human labeling; (2) instead of object tracking, the method integrates particles and social force as feature descriptors which well adapted in both crowded or few people scenes; (3) z-score is used in measuring the normality of events. Due to computation simplicity, the normality detection could be real-time once the training is finished.

目錄
第一章 緒論	1
第二章 相關文獻回顧	4
第三章 理論基礎及相關模型簡介	8
3.1 光流法 (Optical Flow)	8
3.2 社會力模型 (Social Force Model)	10
3.3 Cuboid	11
3.3 Histogram of Oriented Gradient	13
3.4 K-Means	15
3.5 K-Means++	16
第四章 擁擠行為之異常事件偵測系統	18
4.1 系統流程概述	18
4.2 影像中粒子之光流特徵擷取	20
4.3 特徵描述元HOSF建立	21
4.3.1 粒子Social Force 計算	21
4.3.2 各Cuboid之HOSF建立	22
4.4 字典的訓練與建立	24
4.5 異常事件偵測	25
第五章 實驗結果與討論	30
第六章 結論	37
參考文獻	38
附錄：英文論文	43

圖目錄
圖1. 應用監控攝影機自動監測得到的異常行為範例	2
圖2. PYRAMIDAL L-K方法之示意圖	9
圖3. 粒子根據平均光流場的移動變化和其所對應的互動力[6]；左圖為其中一小部份粒子之移動軌跡；右圖為計算出的粒子之互動力。	11
圖4. CUBOID建立之示意圖[21]	12
圖5. HOG描述元擷取之舉例[28]	14
圖6. K-MEANS演算法之舉例及示意圖[30]；(A)為輸入資料；(B)為初始3個SEED選擇及初次分群結果；(C)和(D)為疊代運算過程及更新的分群和群中心；(E)為K-MEANS演算法最終收斂之分群結果。	16
圖7. 異常行為偵測之系統流程圖	19
圖8. 3D網格粒子分割示意圖	20
圖9. 粒子與鄰居粒子所產生SOCIAL FORCE之舉例	22
圖10. SUB-CUBOIDS 切割示意圖	23
圖11. SUB-CUBOID所對應之8-BIN HISTOGRAM	23
圖12. Z-SCORE分佈圖[33]	26
圖13. SEQUENTIAL MEDIAN FILTER方法示意圖	28
圖14. SEQUENTIAL MEDIAN FILTER方法於連續事件處理之舉例	29
圖15. UMN資料庫[32]中三個場景的實例及影片中原始之異常標籤；由上而下分別以場景1、2、3表示；左列為正常事件，右列為異常事件。	31
圖16. 場景2之測試片段；(A)–(F)各時間點T之偵測結果，左上角為異常性標籤；(G)測試片段之時間條及偵測結果，T = 99時人群開始驚慌逃跑，直到T = 127時人群逐漸減少；其中綠色為正常事件(NORMAL)，紅色為異常事件(ABNORMAL)。	32
圖17. 場景2之比較結果	33
圖18. 場景1之測試片段與結果；(A)-(E)各時間點T之偵測結果。	34
圖19. 場景3之測試片段與結果；(A)-(C)各時間點T之偵測結果。	35
圖20. MEHRAN ET AL. 對場景3之事件偵測[6]	36

[1]	T.B. Moeslund, A. Hilton and V. Kru‥ger, “A Survey of Advances in Vision-Based Human Motion Capture and Analysis,” Computer Vision and Image Understanding, vol. 104, no. 2, pp. 90–126, 2006.
[2]	P.K. Turaga, R. Chellappa, V.S. Subrahmanian and O. Udrea, “Machine Recognition of Human Activities: A Survey,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 18, no. 11, pp. 1473–1488, 2008.
[3]	O.P. Popoola and K. Wang, “Video-Based Abnormal Human Behavior Recognition—A Review,” IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews, vol. 42, pp. 6, 865–787, 2012.
[4]	J. Yin, Q. Yang and J.J. Pan, “Sensor-Based Abnormal Human-Activity Detection,” IEEE Transactions on Knowledge and Data Engineering, vol. 20, no. 8, pp. 1082–1090, 2008.
[5]	B. Zhao, L. Fei-Fei and E.P. Xing, “Online Detection of Unusual Events in Videos via Dynamic Sparse Coding,” Computer Vision and Pattern Recognition, pp. 3313–3320, 2011.
[6]	R. Mehran, A. Oyama and M. Shah, “Abnormal Crowd Behavior Detection using Social Force Model,” Computer Vision and Pattern Recognition, pp. 935–942, 2009.
[7]	G.R. Bradski, “Computer Vision Face Tracking for Use in a Perceptual User Interface,” IEEE Workshop on Applications of Computer Vision, pp. 214–219, 1998.
[8]	N. Dalal and B. Triggs, “Histograms of Oriented Gradients for Human Detection,” Computer Vision and Pattern Recognition, vol. 1, pp. 886–893, 2005.
[9]	P. Viola, M.J. Jones and D. Snow, “Detecting Pedestrians using Patterns of Motion and Appearance,” International Journal of Computer Vision, vol. 63, no. 2, pp. 153–161, 2005.
[10]	S. Saxena, F. Bremond, M. Thonnat and R. Ma, “Crowd Behavior Recognition for Video Surveillance,” International Conference on Advanced Concepts for Intelligent Vision Systems, pp. 970–891, 2008.
[11]	A. Adam, E. Rivlin, I. Shimshoni and D. Reinitz, “Robust Real Time Unusual Event Detection using Multiple Fixed-Location Monitors,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 30, no. 3, pp. 555–560, 2008.
[12]	L. Kratz and K. Nishino, “Anomaly Detection in Extremely Crowded Scenes using Spatio-Temporal Motion Pattern Models,” Computer Vision and Pattern Recognition, pp. 1446–1453, 2009.
[13]	V. Mahadevan, W. Li, V. Bhalodia and N. Vasconcelos, “Anomaly Detection in Crowded Scenes,” Computer Vision and Pattern Recognition, pp. 1975–1981, 2010.
[14]	V. Reddy, C. Sanderson and B.C. Lovell, “Improved Anomaly Detection in Crowded Scenes via Cell-based Analysis of Foreground Speed, Size and Texture,” IEEE Workshop on Computer Vision and Pattern Recognition, pp. 55–61, 2011.
[15]	D. Helbing and P. Molnar, “Social Force Model for Pedestrian Dynamics,” Physical Review E, vol. 51, no. 5, pp. 4282–4286, 1995.
[16]	Y. Zhang, L. Qin, H. Yao and Q. Huang, “Abnormal Crowd Behavior Detection Based on Social Attribute-Aware Force Model,” IEEE International Conference on Image Processing, pp. 2689 –2692, 2012.
[17]	R. Raghavendra, A.D. Bue, M. Cristani and V. Murino, “Optimizing Interaction Force for Global Anomaly Detection in Crowded Scenes,” IEEE Workshop on Modeling, Simulation and Visual Analysis of Large Crowds, pp. 136–143, 2011.
[18]	M. Luber, J.A. Stork, G.D. Tipaldi and K.O. Arras, “People Tracking with Human Motion Predictions from Social Forces,” International Conference on Robotics and Automation, pp. 464–469, 2010.
[19]	H. Wang, M.M. Ullah, A. Klaser, I. Laptev and C, Schmid, “Evaluation of Local Spatio-Temporal Features for Action Recognition,” British Machine Vision Conference, pp. 1–11, 2009.
[20]	L. Laptev and T. Lindeberg, “Space-Time Interest Points,” International Conference on Computer Vision, pp. 432–439, 2003.
[21]	P. Dollar, V. Rabaud, G. Cottrell and S. Belongie, “Behavior Recognition via Sparse Spatio-Temporal Features,” IEEE International Workshop on Visual Surveillance and Performance Evaluation of Tracking and Surveillance, pp. 65–72, 2005.
[22]	J.J. Gibson, “The Perception of the Visual World,” Houghton Mifflin, 1950.
[23]	B.K.P. Horn and B.G. Schunck, “Determining Optical Flow,” Artificial Intelligence, vol. 17, pp. 185–203, 1981.
[24]	B.D. Lucas and T. Kanade, “An Iterative Image Registration Technique with an Application to Stereo Vision,” International Joint Conference on Artificial Intelligence, vol. 2, pp. 121–130, 1981.
[25]	J.-Y. Bouguet, “Pyramidal Implementation of the Lucas Kanade Feature Tracker Description of the Algorithm,” Intel Corporation, Microprocessor Research Labs, 2000.
[26]	S.S. Beauchemin and J.L. Barron, “The Computation of Optical Flow,” ACM New York, USA, vol. 27, no. 3, pp. 433–466, 1995.
[27]	C. Cortes and V. Vapnik, “Support-Vector Networks,” Machine Learning, vol. 20, no. 3, pp. 273–297, 1995.
[28]	S.E. Lee, K. Min and T. Suh, “Accelerating Histograms of Oriented Gradients Descriptor Extraction for Pedestrian Recognition,” Computers and Electrical Engineering, vol. 39, issue 4, pp. 1043–1048, 2013.
[29]	J.B. MacQueen, “Some Methods for Classification and Analysis of Multivariate Observations,” Proceedings of 5-th Berkeley Symposium on Mathematical Statistics and Probability, pp. 281–297, 1967.
[30]	A.K. Jain, “Data Clustering: 50 Years beyond K-Means,” Pattern Recognition Letters, vol. 31, issue 8, pp. 651–666, 2010.
[31]	D. Arthur and S. Vassilvitskii, “K-Means++: The Advantages of Careful Seeding,” Proceedings of the 18-th Annual ACM-SIAM Symposium on Discrete Algorithms, pp. 1027–1035, 2007.
[32]	“Unusual Crowd Activity Dataset of University of Minnesota,” available from http://mha.cs.umn.edu/Movies/Crowd-Activity-All.avi.
[33]	Modify from wiki: https://zh.wikipedia.org/wiki/File:Normal_distribution_and_scales.gif