在動態攝影環境下，根據時間整合之行人偵測

簡易檢索 / 詳目顯示

回結果列表

研究生：	吳奇峻 Chi-Jiunn Wu
論文名稱：	在動態攝影環境下，根據時間整合之行人偵測 Temporally Integrated Pedestrian Detection in Non-Static Camera Environment
指導教授：	賴尚宏 Shang-Hong Lai
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：	2006
畢業學年度：	94
語文別：	英文
論文頁數：	49
中文關鍵詞：	行人偵測、動態攝影環境
外文關鍵詞：	Pedestrian Detection, Non-static Camera Environment
相關次數：	點閱：2 下載：0
分享至:	分享至facebook 分享至twitter

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

在這篇論文中，我們提出一個在動態攝影環境下並經由時間上的整合去偵測行人。這套系統主要可分成三個部分，分別為移除背景藉由整體運動估測、利用AdaBoost演算法去偵測行人以及經由時間整合去提升準確率。在移除背景方面，根據一些可靠的區塊和RANSAC演算法去建立一個簡易的Affine整體運動模型。將背景移除之後，我們根據較有可能的前景部分去偵測行人。我們使用了AdaBoost機器學習演算法去針對單張影像找出行人的可能所在位置。最後，經由時間上的資訊，先將每一個現在偵測出來是行人的方框去建立一個圖形的結構，再藉由最佳連線演算法去把相似的方框給聚集成同一個群組。而對於比較大的群組，再把一些中間消失的行人偵測給補救回來。
我們在清華大學裡面實地拍攝了三段影片去測試了系統的效能。實驗結果證明我們的提出的系統可以達到相當高的偵測率和很低的錯誤偵測比率。

In this thesis, we propose an integrated approach that can detect pedestrian system from video sequence acquired with non-static camera environment. The proposed system contains three major components, including global motion estimation with background subtraction, AdaBoost pedestrian detection, and temporal integration. The global motion estimation with background subtraction can reduce the influence of the background pixels and improve the detection accuracy. The simplified affine model is used to fit the global motion model from some reliable blocks by using the RANSAC robust estimation algorithm. After motion-compensated background subtraction, the AdaBoost learning algorithm is employed to detection the pedestrian in a single frame. At last, the graph structure is applied to model the relationship of different detection windows in the temporal domain. The similar detection windows will be grouped as the same clusters by using the optimal linking algorithm. The missed detection windows will be recovered in the clusters containing larger nodes.
In the experimental results, we capture three kinds of video in the campus of National Tsing Hua University to evaluate the performance of our system. The experimental results demonstrate that the proposed system achieves high detection accuracy and low false alarm rate.

List of Figures iii
List of Tables v
Introduction 1
Previous Works 3
1 Image-based Human Detection . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1 Feature Representation using Entire Image . . . . . . . . . . . . . . 4
1.2 Feature Representation using Sub-Image . . . . . . . . . . . . . . . . 7
2 Video-based Pedestrian Detection . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1 Appearance Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Motion Pattern Model . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Proposed Method 12
1 Preliminary Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.1 Robust Estimator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.2 Machine Learning Technique . . . . . . . . . . . . . . . . . . . . . . 15
2 Global Motion Estimation and Background Removal . . . . . . . . . . . . . 16
2.1 Uncertain Parts Rejection . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 A±ne Model Fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3 RANSAC Robust Estimation . . . . . . . . . . . . . . . . . . . . . . 21
3 Pedestrian Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1 Feature Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2 Adaboost Classi‾er . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4 Temporal Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.1 Graph Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.2 Optimal Linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.3 Recover Missed Detection . . . . . . . . . . . . . . . . . . . . . . . . 32
Experiments Results 34
1 Performance of the AdaBoost pedestrian detector . . . . . . . . . . . . . . . 35
ii
1.1 Performance Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.2 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2 Video Sequences in campus . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Conclusion and Future works 43
Bibliography 45

List of Figures
1 A example of 2-D Harr wavelet templates: (a) the vertical wavelets, (b) the horizontal wavelets, (c) the diagonal wavelets . . . . . . . . . . . . . . . . . 4
2 Examples of silhouette Templates . . . . . . . . . . . . . . . . . . . . . . . . 6
3 Examples of the nonhuman edge pixels. (a) the original images (b) the edge pixels of the images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1 The flow chart of proposed method . . . . . . . . . . . . . . . . . . . . . . . 14
2 A illustration of ambiguous of uncertain parts: (a) the real motion vectors between current frame and reference frame, (b) the block A estimate the correct motion filed, but the block B and C will get confused in computing the motion field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3 The flow chart of the proposed global motion estimation with background removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4 Example I of the homogeneous and 1D structure blocks: (a) the original image, (b) the uncertain blocks . . . . . . . . . . . . . . . . . . . . . . . . . 19
5 Example II of the homogeneous and 1D structure blocks: (a) the original image, (b) the uncertain blocks . . . . . . . . . . . . . . . . . . . . . . . . . 20
6 Example I of the global motion estimation: (a) the original image, (b) the results of diamond motion estimation, (c) the global motion estimation . . . 22
7 Example II of the global motion estimation: (a) the original image, (b) the results of diamond motion estimation, (c) the global motion estimation . . . 23
8 Examples of the foreground energy: (a), (b), and (c) are the original images. (d), (e), and (f) are the foreground energy with respect to the global motion model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
9 The flow chart of pedestrian detection system . . . . . . . . . . . . . . . . . 25
10 The °ow chart of temporal integration system . . . . . . . . . . . . . . . . . 28
11 The link of graphic model. ft stands for the tth frame. Assuming the ft~ft+3 has one person in it. The blue nodes present the same person in this sequence. (a) the linking between two consecutive frames, (b) the linking between nearby W frames. W is set to 2 in this case. . . . . . . . . . . . . . 28
12 A illustration of the existed approaches to solve the graphic model. This sequences contain two di®erent persons marked as blue nodes and red nodes. (a) the ground truth of the graphic model, (b) the greedy algorithm to search the closet nodes, (c) the minimum-cut/maximum-flow algorithm . . . . . . 30
13 A example of recovering the missed detection . . . . . . . . . . . . . . . . . 33
1 The examples of positive training data: (a) the INRIA database, (b) the MIT database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2 The examples of negative training data . . . . . . . . . . . . . . . . . . . . . 37
3 The detection for frame 107th-109th in results of Sequence 1 by using (a) AdaBoost detection algorithm only, (b) AdaBoost detection + Background removal, and (c) the proposed pedestrian detection method . . . . . . . . . 39
4 The detection for frame 448th-450th in results of Sequence 1 by using (a) AdaBoost detection algorithm only, (b) AdaBoost detection + Background removal, and (c) the proposed pedestrian detection method . . . . . . . . . 39
5 the pedestrian detection results of applying the proposed algorithm to sequence 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6 the pedestrian detection results of applying the proposed algorithm to sequence 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7 the pedestrian detection results of applying the proposed algorithm to sequence 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

List of Tables
1 The MIT and INRIA human databases for training and testing of our pedestrian detection algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2 Experimental comparison for the pedestrian detection based on SVM and AdaBoost classi‾ers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

List of Algorithms
RANdom SAmple Consensus Algorithm . . . . . . . . . . . . . . . . . . . . 15
Adaptive RANSAC Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 16
AdaBoost Learning Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 17
Global Motion Modeling based on Adaptive RANSAC algorithm . . . . . . 22
Pedestrian Detection based on modified AdaBoost Training Algorithm . . . 27
The algorithm of Optimal Linking . . . . . . . . . . . . . . . . . . . . . . . 32

                                

[1] Shivani Agarwal and Dan Roth. Learning a sparse representation for object detection. In Proceedings of the European Conference on Computer Vision, pages 113-130, London, UK, 2002.
[2] Shai Avidan. Support vector tracking. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, volume 1, pages 184-191, 2001.
[3] Shai Avidan. Support vector tracking. IEEE Trans. on Pattern Analysis and Machine Intelligence, 26:1064-1072, 2004.
[4] MICHAEL J. BLACK and P. ANANDAN. The robust estimation of multiple motions: Parametric and piecewise-smooth flow fields. Computer Vision and Image Understanding, 63:75-104, 1996.
[5] Hwann-Tzong Chen, Horng-Horng Lin, and Tyng-Luh Liu. Multi-object tracking using dynamical graph matching. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, volume 2, pages 210-217, 2001.
[6] Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest, and Cli®ord Stein. Introduction to Algorithms. 2001.
[7] Navneet Dalal and Bill Triggs. Histograms of oriented gradients for human detection. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, volume 1, pages 886-893, 2005.
[8] M. A. Fischler and R. C. Bolles. Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography. Communications of the ACM, 24(6):381-395, 1981.
[9] Yoav Freund and Robert E. Schapire. A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences, 55:119-139, 1996.
[10] D. Gavrila. Pedestrian detection from a moving vehicle. In Proceedings of the 6th European Conference on Computer Vision, pages 37-49, London, UK, 2000.
[11] D. M. Gavrila and V. Philomin. Real-time object detection for smart vehicles. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pages 87-93, Fort Collins Colorado USA, 1999.
[12] Ismail Haritaoglu, David Harwood, and Larry S. Davis. W4: Real time surveillance of people and their activities. IEEE Trans. on Pattern Analysis and Machine Intelligence, 22:809-830, 2000.
[13] C. Harris and M. Stephens. A combined corner and edge detector. In in Alvey Vision Conference, pages 147-151, 1988.
[14] Richard Hartley and Andrew Zisserman. Multiple View Geometry in Computer Vision. 2000.
[15] Bastian Leibe, Ales Leonardis, and Bernt Schiele. Combined object categorization and segmentation with an implicit shape model. In Proceedings of the European Conference on Computer Vision Workshop on Statistical Learning in Computer Vision, pages 17-32, 2004.
[16] Bastian Leibe and Bernt Schiele. Interleaved object categorization and segmentation. In in British Machine Vision Conference, pages 759-768, Norwich, UK, 2003.
[17] Bastian Leibe, Edgar Seemann, and Bernt Schiele. Pedestrian detection in crowded scenes. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, volume 1, pages 878-885, 2005.
[18] Stan Z. Li and ZhenQiu Zhang. Floatboost learning and statistical face detection. IEEE Trans. on Pattern Analysis and Machine Intelligence, 26:1112-1123, 2004.
[19] Anuj Mohan, Constantine Papageorgiou, and Tomaso Poggio. Example-based object detection in images by components. IEEE Trans. on Pattern Analysis and Machine Intelligence, 23(4):349-361, 2001.
[20] C. Papageorgiou, M. Oren, and T. Poggio. A general framework for object detection. In Proceedings of the IEEE International Conference on Computer Vision, pages 555-562, 1998.
[21] Constantine Papageorgiou and Tomaso Poggio. A trainable system for object detection. International Journal of Computer Vision, 38(1):15-33, 2000.
[22] V. de Poortere, J. Cant, B. Van den Bosch, J. dePrins, F. Fransens, and L. Van Gool. Efficient pedestrian detection: a test case for svm based categorization. In Workshop on Cognitive Vision, 2002.
[23] Hedvig Sidenbladh. Detecting human motion with support vector machines. In IAPR International Conference on Pattern Recognition, volume 2, pages 188-191, Cambridge, UK, 2004.
[24] Chris Stau®er and W. Eric L. Grimson. Learning patterns of activity using real-time tracking. IEEE Trans. on Pattern Analysis and Machine Intelligence, 22:747-757, 2000.
[25] Ying-Li Tian, Max Lu, and Arun Hampapur. Robust and efficient foreground analysis for real-time video surveillance. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, volume 1, pages 1182-1187, 2005.
[26] Paul Viola and Michael J. Jones. Robust real-time face detection. International Journal of Computer Vision, 57(2):137-154, 2004.
[27] Paul Viola, Michael J. Jones, and Daniel Snow. Detecting pedestrians patterns of motion and appearance. In Proceedings of IEEE International Conference on Computer Vision, volume 2, pages 734-741, 2003.
[28] Ying Wu and Ting Yu. A ‾eld model for human detection and tracking. IEEE Trans. on Pattern Analysis and Machine Intelligence, 28:753-765, 2006.
[29] Ying Wu, Ting Yu, and Gang Hua. A statistical field model for pedestrian detection. In Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, volume 1, pages 1023-1030, 2005.
[30] S. Zhu and K. K. Ma. A new diamond search algorithm for fast block-matching motion estimation. IEEE Trans. On Image Processing, 9:287-290, 2000.

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)

簡易檢索 / 詳目顯示

相關論文