臺灣博碩士論文加值系統

不論在安全監控、人機互動、電腦動畫或甚至醫學應用上，人物動作的擷取與分析是個相當重要的議題。在本論文中，我們提出一個在多攝影機環境下，利用人體模型估測目標人物的姿勢與行為。我們使用流形嵌入技術中的拉普拉斯特徵映射，將三維人形的幾何形狀忠實地轉移到另一個容易切割分析的高維度空間，正確地切割出三維人形的各個部位並且找出三維人形的骨骼架構，以利後續行為分析的動作。當擷取出三維人形的骨骼架構後，我們利用粒子群體最佳化在高維度空間中有效地找出最佳姿態估測結果。我們的系統由影像的擷取至姿態的估測完全自動化，並且不需要在人體上貼附感應物，即可結合肢體的運動限制和時間軸上的動作流暢限制，估測多種動作。

Chapter 1. Introduction 1
Chapter 2. Backgrounds 3
2.1 Motion Capture Systems 3
2.1.1 Model-Based Systems 4
2.1.1.1 Monocular Motion Capture 4
2.1.1.2 Multiple View Motion Capture 5
2.1.2 Model-Free Systems 10
2.2 Manifold Learning 13
2.2.1 Manifold Embedding Methods 13
2.2.2 Embedding Based Initialization 15
Chapter 3. Proposed Method 20
3.1 Initialization of Motion Capture 21
3.1.1 Images Acquisition 21
3.1.2 Camera Calibration 22
3.1.3 Visual Hull Reconstruction 22
3.1.4 Human Body Segmentation 30
3.1.4.1 Sunaresan’s Method 30
3.1.4.2 Modified Body Parts Segmentation Method 34
3.1.5 Skeleton Extraction 40
3.1.5.1 Sundaresan’s Method 40
3.1.5.2 Modified Skeleton Extraction Method 41
3.2 PSO Based Pose Estimation 41
3.2.1 3-D Human Model 42
3.2.1.1 Twists and Exponential Products Formulation 42
3.2.1.2 3-D Skeleton Model 44
3.2.2 PSO Based Pose Estimation 49
3.2.2.1 Evaluation Function 49
3.2.2.2 Hierarchical PSO Fitting Process 51
3.2.2.3 Motion Constraints 53
3.2.2.4 Fine Tune the Pose Estimation Results 54
3.2.2.5 Temporal Consistency 55
Chapter 4. Experimental Results 57
Chapter 5. Conclusions 75
Reference 76

[1] http://www.math.umn.edu/~wittman/mani/
[2] http://mathworld.wolfram.com/Manifold.html
[3] A. Agarwal and B. Triggs, “3-D Human Pose from Silhouettes by Relevance Vector Regression,” IEEE Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 882-888, June, 2004.
[4] A. Elgammal, and C. S. Lee, “Inferring 3-D Body Pose from Silhouettes Using Activity Manifold Learning, ” IEEE Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 681-688, June, 2004.
[5] A. Laurentini, “How Many 2-D Silhouettes Does It Take to Reconstruct a 3-D Object?” Computer Vision and Image Understanding, vol. 67, no.1, pp. 81-87, 1997.
[6] A. Sundaresan, and R. Chellappa, “Model Driven Segmentation of Articulating Humans in Laplacian Eigenspace,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007.
[7] C. Menier, E. Boyer, and B. Raffin,”3-D Skeleton-based Body Pose Recovery,” International Symposium on 3-D Data Processing, Visualization and Transmission, pp. 389-396, 2006.
[8] C. Robertson and E. Trucco, “Human Body Posture via Hierarchical Evolutionary Optimization,” British Machine Vision Conference, vol. 3, pp. 999-1008, 2006.
[9] C. W. Chu, O. C. Jenkins, and M. J. Mararić, “Markerless Kinematic Model and Motion Capture from Volume Sequences,” IEEE Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 475-482, June, 2003.
[10] F. Cuzzolin, D. Mateus, E. Boyer, and R. Horaud, “Robust Spectral 3-D-Bodypart Segmentation along Time,” International Conferences on Computer Vision 2nd Workshop on Human Motion, pp. 196-211, 2007.
[11] G. R. Taylor, A. J. Chosak, and P. C. Brewer, “OVVV: Using Virtual Worlds to Design and Evaluate Surveillance Systems, ”IEEE Conference on Computer Vision and Pattern Recognition, pp. 1-8, June 2007.
[12] I. H. Chen and S J. Wang, “Efficient Vision-Based Calibration for Visual Surveillance Systems with Multiple PTZ Cameras”, in Proceeding of IEEE International Conference on Computer Vision Systems, Jan. 5-7, 2006.
[13] I. Mikic, M. Trivedi, E. Hunter, and P. Cosman, “Human Body Model Acquisition and Tracking Using Voxel Data,” International Journal of Computer Vision, vol.53, pp. 199-223, 2003.
[14] J. B. Tenenbaum, V. D. Silva, and J. C. Langford, “A Global Geometric Framework for Nonlinear Dimensionality Reduction,” Science, vol. 290, pp. 2319-2323, Dec., 2000.
[15] J. Deutscher, A. Blake, and I. Reid, “Articulated Body Motion Capture by Annealed Particle Filtering,” Computer Vision and Image Understanding, vol. 2, pp. 126-133, June, 2000.
[16] J. Kennedy and R. Eberhart, “Particle Swarm Optimization,” Proceedings of the IEEE International Conference on Neural Networks, vol. 4, pp. 1942-1948, 1995.
[17] L. Mündermann, S. Corazza, A. M. Chaudhari, E. J. Alexander, T. P. Andriacchi, “Most Favorable Camera Configuration for a Shape-from-Silhouette Markerless Motion Capture System for Biomechanical Analysis,” Proceedings of Society of Photo-Optical Instrumentation Engineers, pp. 278-287, Jan., 2005.
[18] M. Belkin and P. Niyogi, “Laplacian Eigenmaps for Dimensionality Reduction and Data Representation,” Neural Comput., pp. 1373-1396, 2003.
[19] M. Piccardi and T. Jan, “Mean-shift background image modeling,” International Conference on Image Processing, vol. 5, pp. 3399–3402, Oct. 2004
[20] N. R. Howe, “Silhouette Lookup for Automatic Pose Tracking,” Proceedings of the Conference on Computer Vision and Pattern Recognition Workshops, pp. 15-22, June, 2004.
[21] R. Kehl, L. V. Gool, “Markerless Tracking of Complex Human Motions from Multiple Views,” Computer Vision and Image Understanding, vol. 104, pp. 190-209, 2006.
[22] R. Kehl, M. Bray, and L.V. Gool, “Full Body Tracking from Multiple Views Using Stochastic Sampling,” IEEE Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 129-136, June, 2005.
[23] R. Murray, Z. Li, S. Sastry, “A Mathematical Introduction to Robotic Manipulation,” CRC Press, 1993.
[24] R. Szeliski, “Rapid Octree Construction from Image Sequences, ” CVGIP: Image Understanding, vol. 58, pp. 23-32, 1993.
[25] S. T. Roweis and L. K. Saul, “Nonlinear Dimensionality Reduction by Locally Linear Embedding,” Science, vol. 290, pp. 2323-2326, Dec., 2000.
[26] T. B. Moeslund, A. Hilton, and V. Krüger, “A Survey of Advances in Vision-based Human Motion Capture and Analysis,” Computer Vision and Image Understanding, vol. 104, pp. 90-126, 2006.
[27] Y. Sagawa, M. Shimosaka, T. Mori, and T. Sato, “Fast Online Human Pose Estimation via 3-D Voxel Data,” Proceedings of IEEE Conference on Intelligent Robots and Systems, pp. 1034-1040, 2007.