成功大學博碩士論文系統

簡易檢索 / 詳目顯示

回結果列表

研究生：	朱靜璇 Chu, Ching-Hsuan
論文名稱：	使用反向合成式動態外觀模型與因式分解法重建2D-3D形狀與運動軌跡 2D-3D Shape and Motion Recovery using Inverse Compositional Active Appearance Model and Factorization
指導教授：	連震杰 Lien, Jenn-Jier
學位類別：	碩士 Master
系所名稱：	電機資訊學院 - 資訊工程學系 Department of Computer Science and Information Engineering
論文出版年：	2015
畢業學年度：	103
語文別：	英文
論文頁數：	79
中文關鍵詞：	人臉特徵點偵測與追蹤、最佳形狀搜索、直接結合模型、反向式動態外觀模型、因式分解法
外文關鍵詞：	Facial Feature Point Detection and Tracking, Shape Optimized Search, Direct Combined Model, Inverse Compositional Active Appearance Model, Factorization Method
相關次數：	點閱：110 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

本論文提出一套系統，在找到初步的人臉特徵點資訊後，快速追蹤人臉上的特徵點，並利用二維人臉運動的資訊，進一步重建出三維人臉模型。本系統共分為三個部分：1)二維人臉特徵點偵測 2)二維人臉特徵點偵測錯誤修正與追蹤 3)三維人臉模型重建。本系統第一部分的輸入為攝影機拍攝的連續人臉影像，並使用人臉偵測系統找到人臉的位置，確認人臉位置後，結合最佳形狀搜索及直接結合模型兩種演算法，找出初步的人臉特徵點位置。有了特徵點之初始位置後，進入本系統第二部分。本系統第二部分的輸入為第一部分所偵測出之特徵點位置，利用反向式動態外觀模型，計算正面的人臉影像，並藉由分析正面的人臉影像與訓練好的正面人臉影像所產生的誤差，來更新特徵點位置，進而實現特徵點錯誤位置修正與追蹤。找出連續人臉影像之特徵點後，進入系統第三部分。本系統第三部分的輸入為第二部分修正與追蹤後的人臉特徵點位置，利用因式分解法，分解正投影模型，以重建出三維人臉模型。

This dissertation proposes a real-time system to quickly track human facial features and further build a 3D facial shape model with its 2D motion images. This system is composed of three major elements. Part (1) Detections of 2D facial features, Part (2) Error correction and tracking of 2D facial features, Part (3) Constructions of a 3D human facial model. Sequential video frames of a human face are inputs to Part (1). With a human face detection, the face can be located. After a process of positioning the face, by integrating two algorithms, Shape Optimized Search (SOS) and Direct Combined Model (DCM), locations of feature points of the face are roughly found. Then, Part (2) is initiated by receiving the rough locations of the feature points computed from Part (1). By using Inverse Compositional Active Appearance Model (ICAAM), an image of a front face is processed. The facial feature point locations are rectified with analyses of differences between real-time captured image and training image data of the front face. Therefore, error correction and tracking can be achieved. After all feature points from the sequential video frames are identified in Part (2), Part (3) is started by obtaining outputs of Part (2). An orthographical projection model of the face can be formed with a factorization method to construct a 3D shape model from the sequential video frames in 2D motion.

Content
摘要....................................................IV
Abstract.................................................V
Acknowledgement.........................................VI
Content................................................VII
Content of Figure.......................................IX
Content of Table.......................................XII
Chapter 1.	 Introduction............................1
1.1	Motivation.......................................1
1.2	Related Works....................................3
1.3	System Flowchart.................................7
1.4	Thesis Architecture.............................10
Chapter 2.   2D Facial Feature Points Detection.........12
2.1  Using AdaBoost to detect human face................13
2.2  Using Shape Optimized Search to detect initial facial feature points..........................................15
2.2.1  Training Process.................................16
2.2.2  Testing Process..................................19
2.3  2D Facial Feature Point Expansion using Direct Combined Model...................................................24
2.3.1 Direct Combined Model Training Process............25
2.3.2  Direct Combined Model Testing Process............29
Chapter 3.   2D Facial Feature Points Detection Correction and Tracking using Inverse Compositional Active Appearance Models	........................................................32
3.1 Active Appearance Model (AAM).......................33
3.2 Inverse Compositional Active Appearance Model (ICAAM)..36	
3.3 Apply ICAAM to 2D Feature Points Detection Correction  and Tracking................................................38
3.3.1 ICAAM Training Process............................39
3.3.2 ICAAM Testing Process.............................44
Chapter 4.   3D Facial Shape and Motion Recovery using Rotation- and Shape Basis-Constrained Factorization.....48
4.1  3D Rigid Shape and Motion Recovery using Rotation-Constrained Factorization...............................50
4.2  3D Non-Rigid Shape and Motion Recovery using Rotation- and Shape Basis-Constrained Factorization...................55
Chapter 5.   Experimental Results.......................59
5.1  The accuracy of 2D facial feature point detection and error correction........................................61
5.2  Error Analysis.....................................67
5.3  Tracking Result of 2D Facial Feature Point Detection and Tracking System under Pose Variation....................71
5.4  Experimental Result of 3D Facial Shape and Motion Recovery Subsystem...............................................73
Chapter 6.   Conclusion and Future Works................75
Reference...............................................76

                                    

[1] S. Baker, R. Gross, I. Matthews, and T. Ishikawa, "Lucas-Kanade 20 Years on: A Unifying Framework: Part 2, Technical Report CMU-RI-TR-03-01," Carnegie Mellon University Robotics Institute, 2003.
[2] S. Baker and I. Matthews, "Lucas-Kanade 20 Years on: A Unifying Framework: Part 1: The Quantity Approximated, the Warp Update Rule, and the Gradient Descent Approximation," International Journal of Computer Vision, 56(3):221-225, March 2004.
[3] S. Baker, I. Matthews, J. Xiao, R. Gross, T. Kanade and T. Ishikawa, "Real-Time Non-rigid Driver Head Tracking for Driver Mental State Estimation," Proceedings 11th World Congress Intelligent Transportation Systems, 2004.
[4] T.F. Cootes and C.J. Taylor, "Active Shape Models," In 3rd British Machine Vision Conference, pp. 266-275, 1992.
[5] C. Cortes and V. Vapnik, "Support-Vector Networks," Machine Learning, 1995.
[6] T.F. Cootes, G.J. Edwards, and C.J. Taylor, "Active Appearance Models," In 5th European Conference on Computer Vision, Vol. 2, pp. 484-498, 1998.
[7] T.F. Cootes and C.J. Taylor, "Statistical Models of Appearance for Computer Vision," University of Manchester, 2004.
[8] D. Cristinacce and T. Cootes, "A Comparison of Shape Constrained Facial Feature Detectors," Conference of Automatic Face and Gesture Recognition, pp. 375-380, 2004.
[9] P. Fua, “Regularized Bundle-Adjustment to Model Heads from Image Sequences without Calibration Data,” International Journal of Computer Vision, Vol.38, no.2, pp.153-171, 2000.
[10] R. Gross, I. Matthews, and S. Baker, "Constructing and Fitting Active Appearance Models with Occlusion," Computer Vision and Pattern Recognition Workshop, 2004.
[11] Y. Gong and W. Xu, “Machine Learning for Multimedia Content Analysis,” Springer, 2007.
[12] A.K. Jain, Y. Zhong, and S. Lakshmanan, "Object Matching using Deformable Templates," IEEE Transactions on Pattern Analysis and Machine Intelligence, pp. 267-278, 1996.
[13] Y. C. Lee, “Terzopoulos and K. Waters, Constructing Physics-Based Facial Models of Individuals,” Graphics Interface, pp.1-8, 1993.
[14] S.Z. Li, “Markov Random Field Modeling in Image Analysis,” Springer, 2009.
[15] J. L. Mundy and A. Zisserman, “Geometric Invariance in Computer Vision,” MIT Press, pp.512, 1992.
[16] I. Matthews, J. Xiao, S. Baker, “2D vs. 3D deformable face models: Representational power, construction, and real-time fitting,” International Journal of Computer Vision, 75(1), 93-113, 2007.
[17] M.T. Mahmood, S. Shim and T.S. Choi, “Shape from Focus Using Principal Component Analysis in Discrete Wavelet Transform, Optical Engineering,” Vol.48, no.5, pp.057203:1-057203:9, 2009.
[18] M.S. Muhammad and T.S. Choi, “A Novel method for Shape from Focus in Microscopy using Bezier Surface approximation, Microscopy Research and Technique,” Vol.73, no.2, pp.140-151, 2010.
[19] J.A. Nelder and R. Mead, "A Simplex Method for Function Minimization," Computer Journal, pp. 308-313, 1965.
[20] S. Nayar and N. Yasuo, “Shape From Focus,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.16, no.8, pp.824-831, 1994.
[21] C. Poelman and T. Kanade, “A Paraperspective Factorization Method for Shape and Motion Recovery,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.19, no.3, pp.206-218, 1997.
[22] H. Shum, Q. Ke and Z. Zhang, “Efficient Bundle Adjustment with Virtual Key Frames: A Hierarchical Approach to Multi-Frame Structure from Motion,” Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition, Vol.2, pp.538-543, 1999.
[23] Y. Shan, Z. Liu and Z. Zhang, “Model-Based Bundle Adjustment with Application to Face Modeling,” Proceedings of IEEE International Conference on Computer Vision, pp.644-751, 2001.
[24] M. Turk and A. Pentland, "Eigenfaces for Recognition," Journal of Cognitive Neuroscience, Vol. 3, No. 1, pp. 73-86, 1991.
[25] C. Tomasi and T. Kanade, “Shape and Motion from Image Streams under Orthography: a Factorization Method,” International Journal of Computer Vision, Vol.9, no.2, pp.137-154, 1992.
[26] A. Tankus, N. Sochen and Y. Yeshurun, “A New Perspective on Shape-from-Shading,” Proceedings of IEEE International Conference on Computer Vision, Vol.II, pp.862–869, 2003.
[27] A. Thelen, S. Frey, S. Hirsch and P. Hering, “Improvements in Shape-From-Focus for Holographic Reconstructions With Regard to Focus Operators, Neighborhood-Size, and Height Value Interpolation,” IEEE Transactions on Image Processing, Vol.18, no.1 pp.151-157, 2009.
[28] C. Tu and J.J. Lien, “Direct Combined Model for Facial Feature Point Detection, Face Sketch Synthesis and Occluded Face Recovery,” Ph.D. Thesis, National Cheng Kung University, 2010.
[29] P. Viola and M. Jones, "Rapid Object Detection using a Boosted Cascade of Simple Features," Conference Computer Vision and Pattern Recognition, 2001.
[30] M. Valstar, B. Martinez, X. Binefa, and M. Pantic, "Facial Point Detection using Boosted Regression and Graph Models," Computer Vision and Pattern Recognition, pp. 2729-2736, 2010.
[31] J.Y. Wu and J.J. Lien, “3D Facial Surface Reconstruction Using Integrated Orthographic Models to Approximate Perspective Projection Model,” Ph.D. Thesis, National Cheng Kung University, 2011.
[32] J. Xiao, S. Baker, I. Matthews and T. Kanade, "Real-Time Combined 2D+3D Active Appearance Models," Conference on Computer Vision and Pattern Recognition, 2004.
[33] J. Xiao, J. Chai and T. Kanade, "A Closed-Form Solution to Non-rigid Shape and Motion Recovery," International Journal of Computer Vision, pp. 233-246, 2006.

校內：2025-12-31公開
校外：不公開電子論文尚未授權公開，紙本請查館藏目錄

簡易檢索 / 詳目顯示

相關論文