簡易檢索 / 詳目顯示

回結果列表
研究生： 張朝陽
Chao-Yang Chang
論文名稱： 以軌跡為基礎的監視影像事件偵測及事件發現
Trajectory-Based Event Detection and Discovery for Surveillance Videos
指導教授： 許秋婷
Chiou-Ting Hsu
口試委員:
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 資訊工程學系
Computer Science
論文出版年： 2008
畢業學年度： 96
語文別： 英文
論文頁數： 59
中文關鍵詞： 事件偵測 、事件發現 、軌跡
外文關鍵詞： Event Detection, Event Discovery, Trajectory
相關次數： 點閱：2  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 在影像監視系統(video surveillance system)中的事件偵測是很要的一個部分，因為人們通常會對某些事件感到興趣並且想要去偵測他們。這篇論文提出了一個偵測事件(event detection)的方法。在影像監視系統中軌跡是一種表現物體行為的重要方式，因此我們使用軌跡做為我們事件偵測的特徵。首先，我們使用多個物體追蹤系統(multi-object tracking system)來獲取軌跡，並透過中間階層(mid-level)來表示軌跡。接下來，我們使用隱馬爾可夫模型(hidden Markov model)來模擬和偵測事件。我們用中間層來表示軌跡是因為它可以避免雜訊的干擾而且它可以減低我們在學習事件模型時的計算複雜度。然而，有些軌跡是不容易被分類而且人工分類成本太高。再者當環境改變的時候，也許會產生新的事件。這時我們需要一個方法來自動化的解理這些新事件。因此，我們提出了一個事件發現(event discovery)的方法來發現新事件。我們對不屬於任何已存事件的軌跡做分群並且用新事件的限制來看是否分出的群是新的事件。

    中文摘要 I ABSTRACT II 1. INTRODUCTION 1 2. RELATED WORK 3 2.1 Low-level Based Method 3 2.2 String-matching Based method 4 2.3 HMM Based Method 5 2.4 Segmentation Based method 7 3. DATA REPRESENTATION 11 3.1 Trajectory Extraction 11 3.1.1 Foreground Extraction 11 3.1.2 Multi-object Tracking 15 3.2 Mid-level Feature Extraction 16 3.2.1 Data Segmentation 16 3.2.2 Sub-trajectory Clustering 17 3.2.3 Mid-level Feature Extraction 17 4. EVENT DETECTION AND EVNT DISCOVERY 23 4.1 Event Detection 23 4.1.1 HMM Parameters and HMM Learning 24 4.1.2 Event Detection 24 4.2 Event Discovery 25 4.2.1 Outlier Decision 25 4.2.2 Distance Measurement 26 4.2.3 Clustering Algorithm 27 4.2.4 New Event Constraint 28 5. EXPERIMENTIAL RESULTS 32 5.1 Trajectory Feature 32 5.2 Event Detection 32 5.2.1 Accuracy Measurement 33 5.2.2 Lobby Scene 33 5.2.3 Crossroad Scene 34 5.3 Noise Resistance 35 5.4 Event Discovery 35 5.4.1 Outlier Detection 35 5.4.2 Discovered Event 36 5.4.3 Un-discovered Event 36 5.5 Computational Time 37 5.6 Block Size Decision 37 5.7 Discussion 37 6. CONCLUSIONS 57 7. REFERENCES 58

    [1] S.Y Yang, “A Study of Moving Objects Extraction for Surveillance Videos by Background-Subtraction Method,” Master Thesis, National Tsing Hua University (Taiwan), June 2006.
    [2] Y. Wang, K. F. Loe, and J. K. Wu, “A Dynamic Conditional Random Field Model for Foreground and Shadow Segmentation,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 2, pp. 279-289, Feb. 2006.
    [3] G. Gennari and G.D. Hager, “Probabilistic Data Association Methods in Visual Tracking of Group,” Proc. CVPR’04, vol. 2, pp. 876-881, July 2004.
    [4] Fatih Porikli and Tetsuji Haga, “Event Detection by Eigenvector Decomposition Using Object and Frame Features,” Proc. CVPRW’04, June 2004.
    [5] G. Schwarz, “Estimating the dimension of a model,” Annuals of Statistics, vol. 6, pp. 461-464, 1978.
    [6] L.R. Rabiner, “A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition,” Proc. IEEE, vol. 77, no. 2, pp. 257-286, Feb. 1989.
    [7] Zhouyu Fu, Weiming Hu, and Tienir Tan, “Similarity Based Vehicle Trajectory Clustering and Anomaly Detection,” Proc. ICIP’05, vol. 2, pp. 602-605, Sept. 2005.
    [8] Xi Li, Weiming Hu, and Wei Hu, “A Coarse-to-Fine Strategy for Vehicle Motion Trajectory Clustering, ” Proc. ICPR’06, vol. 1, pp. 591-594, 2006.
    [9] Jiang-Bin Zheng, David Dagan Feng, and Ring-Chun Zhao, “Trajectory Matching and Classification of Video Moving Objects,” Proc. Multimedia Signal Processing, pp. 1-4, Oct. 2005.
    [10] Dan Buzan, Stan Sclaroff, and Gerge Kollios, “Extraction and Clustering of Motion Trajectories in Video,” Proc. ICPR’04, vol. 2, pp. 521-524, Aug. 2004.
    [11] Zhang Zhang, Kaiqi Huang, Tieniu Tan, Liangsheng Wang, “Trajectory Series Analysis based Event Rule Induction for Visual Surveillance,” Proc. CVPR’07, pp. 1-8, June 2007.
    [12] Jacinto C. Nascimento, Mario A. T. Figueiredo, and Jorge S. Marques, “Semi-supervised Learning of Switched Dynamical Models for Classification of Human Activities in Surveillance Application,” Proc. ICIP’07, vol. 3. pp. 197-200, Dept. 2007.
    [13] A.Hervieu, P. Bouthemy, and J-P. Le Cadre, “A HMM-based method for Recognizing Dynamic Video Contents from Trajectories,” Proc. ICIP’07, vol. 4, pp. 533-536, Dept. 2007.
    [14] Y. Wang, K. F. Loe, and J. K. Wu, “A Dynamic Conditional Random Field Model for Foreground and Shadow Segmentation,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 2, pp. 279-289, Feb. 2006.
    [15] Fan Jiang, Ying Wu, and Agglos K. Katsaggelos, “Abnormal Event Detection from Surveillance Video by Dynamic Hierarchical Clustering,” Proc. ICIP’07, vol. 5, pp. 145-148, Sept. 2007.

    無法下載圖示 全文公開日期 本全文未授權公開 (校內網路)
    全文公開日期 本全文未授權公開 (校外網路)

    QR CODE