由於人臉並非剛體，表情及姿態的改變會導致影像有很大的變異，並且還有許多外在因素如光照、遮蔽等影響。因此要達成穩定的追蹤需要有多個不同姿態並且穩健的人臉模型以進行量測。另外，當人臉在追蹤過程可能因遮蔽區域過大而導致丟失，為了能夠尋回追蹤的對象，個人的特徵需事先被學習。然而，如何克服追蹤過程中的干擾，並線上建構出多姿態的個人人臉特徵模型是個挑戰。而線上人臉識別的應用中，如果能夠收集愈充份的資訊，則能夠提升人臉識別的準確率，然而資訊的收集也會因為追蹤過程的穩定性所影響，因此如何收集足夠且正確的資訊也是一個難題。
在本篇論文中，我們提出了一個整合泛用人臉模型及個人人臉模型的追蹤演算法，並利用人臉色彩核長條圖的輔助，以達成穩定的人臉追蹤。其中泛用人臉模型可以協助個人人臉模型的線上學習以建構出多姿態的個人人臉模型。經由多姿態的個人人臉模型，除了可以在追蹤過程中進行目標的對應，也能夠在目標因特殊狀況而丟失時，透過丟失尋回的機制將丟失的目標尋回。由於線上學習的個人人臉模型是利用LBP特徵空間所建構的區域紋理特徵，可以抵抗局部遮蔽，更進一步的，其所學習的資訊可被直接使用於人臉識別。
經由我們的實驗証實，透過整合泛用人臉模型及個人人臉模型的追蹤演算法所建構的多姿態個人人臉模型，可以在複雜背景、光度變化、局部遮蔽、姿態改變等不同因素干擾下完成目標的追蹤，並在人臉追蹤器丟失時能夠正確的找回對應的人臉。而其個人人臉模型所包含的資訊能夠直接被應用於線上人臉識別，並且透過多個人臉模型的整合可以有效提升其人臉識別的準確性。經由測試，其辨識的正確率可達到70%以上。

Because human face is not a rigid object, the changes of face expression or poses would cause huge variation in the image. Furthermore, there are other disturbances such as the varying of illumination and partial occlusions. Therefore, it is necessary to have a robust multi-pose measurement model to achieve stable tracking. In addition, tracked face could be lost when the occluded region is too large. To recover the tracking, specific features are needed to learn previously. However, how to overcome the disturbance and construct the multi-pose specific feature model is a challenge. Regarding to on-line face recognition, the more personal information we collect, the more accurate result we’ll get. The collection of such information would also be affected by the instability of tracking. How to obtain correct information for on-line face recognition is a problem.
In this thesis, we propose an integrated tracking algorithm combining generic face model and specific face model. We use the color kernel histogram of human face to assistant the integration of combining two models, and use generic face model to help construct multi-pose specific face model. Via the specific face model, even if the tracking target is lost in some situations, the model can help to find the losing target. Because the specific face model is constructed by LBP texture feature, it can achieve robust tracking, including partial occlusion. And the learned information of specific face model can be used in face recognition.
In our experiment, the purposed method can achieve good tracking result in the condition of complex background, varying of illumination, partial occlusion, changes of poses and so on. The target losing during tracking can be recovered correctly as well. For face recognition, the multi-pose specific face model can provide sufficient information to achieve acceptable accuracy rate. Through the experiment the accuracy rate is above 70%.

[1] K. K. Sung and T. Poggio, "Example-based learning for view-based human face detection," IEEE Transactions on Pattern Analysis and Machine Intelligence,vol. 20, pp. 39-51, 1998.
[2] H. A. Rowley, S. Baluja, and T. Kanade, "Neural network-based face detection," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 20, pp. 23-38, 1998.
[3] P. Viola and M. J. Jones, "Rapid object detection using a boosted cascade of simple features," In Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. I-511-I-518 vol.1, 2001.
[4] P. Viola and M. J. Jones, "Robust real-time face detection," In Proceedings. Eighth IEEE International Conference on Computer Vision, pp. 747-747, 2001.
[5] P. Viola and M. J. Jones, "Fast multi-view face detection," TR2003-96, Mitsubishi Electric Research Laboratories, August 2003.
[6] J. Friedman, T. Hastie, and R. Tibshirani, "Additive logistic regression: a statitical view of boosting," Ann, Statist., vol. 28, no. 2, pp.337-374,2000.
[7] 謝元澄, "使用聚類來加速 Adaboost 並實現噪聲數據探測," Journal of Software, vol. 21, pp. 1889-1897, 2010.
[8] A. Hadid, J. Y. Heikkila, O. Silven, and M. Pietikainen, "Face and eye detection for person authentication in mobile phones," In First ACM/IEEE International Conference on Distributed Smart Cameras, pp. 101-108, 2007.
[9] C. Demirkir and B. Sankur, "Face detection in cluttered scenes using look-up table based boosting algorithm," In Proceedings of the IEEE 13th Signal Processing and Communications Applications Conference , pp. 460-463, 2005.
[10] M. Isard and A. Blake, "Condensation: conditional density propagation for visual tracking," International Journal of Computer Vision, vol. 29, pp. 5-28, 1998.
[11] M. S. Arulampalam, S. Maskell, N. Gordon, and T. Clapp, "A tutorial on particle filters for online nonlinear/non-gaussian bayesian tracking," IEEE Transactions on Signal Processing, vol. 50, pp. 174-188, 2002.
[12] W. Peng and J. Qiang, "Robust face tracking via collaboration of generic and specific models," IEEE Transactions on Image Processing, vol. 17, pp. 1189-1199, 2008.
[13] Y. Utsumi and Y. Iwai, "Face tracking and recognition by using omnidirectional sensor network," In Third ACM/IEEE International Conference on Distributed Smart Cameras, pp. 1-8, 2009.
[14] R. C. Verma, C. Schmid, and K. Mikolajczyk, "Face detection and tracking in a video by propagating detection probabilities," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, pp. 1215-1228, 2003.
[15] Z. Wenlong and S. M. Bhandarkar, "A boosted adaptive particle filter for face detection and tracking," In 2006 IEEE International Conference on Image Processing, pp. 2821-2824, 2006.
[16] L. Wang, T. Tan, and W. Hu, "Face tracking using motion-guided dynamic template matching," presented at the The 5th Asian Conference on Computer Vision Melbourne, Australia, 2002.
[17] D. Comaniciu, V. Ramesh, and P. Meer, "Real-time tracking of non-rigid objects using mean shift," In Proceedings. IEEE Conference on Computer Vision and Pattern Recognition, pp. 142-149 vol.2, 2000.
[18] L. Kuang-Chih and D. Kriegman, "Online learning of probabilistic appearance manifolds for video-based recognition and tracking," In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 852-859 vol. 1, 2005.
[19] L. Yongmin, G. Shaogang, and H. Liddell, "Support vector regression and classification based multi-view face detection and recognition," In Proceedings. Fourth IEEE International Conference on Automatic Face and Gesture Recognition, pp. 300-305, 2000.
[20] J. Myung-Ho and K. Hang-Bong, "Illumination invariant face tracking and recognition," In 5th European Conference on Visual Media Production (CVMP 2008), pp. 1-6,2008.
[21] R. C. Luo, C. T. Liao, and Y. J. Chen, "Robot - human face tracking and recognition using relative affine structure," In IEEE Workshop on Advanced robotics and Its Social Impacts, pp. 1-6, 2008.
[22] G. R. Bradski, "Computer vision face tracking for use in a perceptual user interface," Intel Technology Journal, vol. 2, pp. 13-27, 1998.
[23] Z. H. Khan, I. Y. Gu, and A. G. Backhouse, "Robust visual object tracking using multi-mode anisotropic mean shift and particle filters," IEEE Transactions on Circuits and Systems for Video Technology , vol. 21, pp. 74-87, 2011.
[24] W. Xian, L. Lihong, L. Jianhuang, and H. Jian, "A framework of face tracking with classification using CAMShift-C and LBP," In Fifth International Conference on Image and Graphics, pp. 217-222, 2009.
[25] X. Du, C. Liu, and Y. Yu, "Analysis of detection and track on partially occluded face," International Forum on Information Technology and Applications, pp. 158-161, 2009.
[26] J. Harguess, H. Changbo, and J. K. Aggarwal, "Occlusion robust multi-camera face tracking," In IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, pp. 31-38, 2011.
[27] W. Ying, Y. Ting, and H. Gang, "Tracking appearances with occlusions," In Proceedings. 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. I-789-I-795 vol.1, 2003.
[28] B. Peng, Q. Hong, W. Anhua, and L. Yu, "Person-tracking with occlusion using appearance filters," In 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 1805-1810,2006.
[29] V. Ngoc-Son and A. Caplier, "Efficient statistical face recognition across pose using local binary patterns and gabor wavelets," In IEEE 3rd International Conference on Biometrics: Theory, Applications, and Systems, pp. 1-5, 2009.
[30] S. Liao, X. Zhu, Z. Lei, L. Zhang, and S. Z. Li, "Learning multi-scale block local binary patterns for face recognition," presented at the International Conference on Biometrics (ICB), 2007.
[31] W. Yunlong, X. Mei, S. Rui, and L. Tao, "Face location with LBP scale transform," In 2010 International Conference on Communications, Circuits and Systems (ICCCAS), pp. 347-350, 2010.
[32] T. Ojala, M. Pietikainen, and T. Maenpaa, "Multiresolution gray-scale and rotation invariant texture classification with local binary patterns," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, pp. 971-987, 2002.
[33] T. Ahonen, A. Hadid, and M. Pietikainen, "Face description with local binary patterns: application to face recognition," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 28, pp. 2037-2041, 2006.
[34] W. Bo, A. Haizhou, H. Chang, and L. Shihong, "Fast rotation invariant multi-view face detection based on real adaboost," In Proceedings. Sixth IEEE International Conference on Automatic Face and Gesture Recognition, pp. 79-84, 2004.
[35] Y. Freund and R. E. Schapire., "A decision-theoretic generalization of on-line learning and an application to boosting," In Computational Learning Theory: Eurocolt '95, pp. 23-37, 1995.
[36] R. E. Schapire and Y. Singer, Improved boosting algorithms using confidence-rated predictions vol. 37: Kluwer Academic, 1999.
[37] L. Zhang, R. Chu, S. Xiang, S. Liao, and S. Z. Li, "Face detection based on multi-block LBP representation," In Proceedings of IAPR/IEEE International Conference on Biometrics, (ICB-2007). Seoul, Korea, August 2007.
[38] D. Comaniciu, V. Ramesh, and P. Meer, "Kernel-based object tracking," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, pp. 564-577, 2003.
[39] M. K. Osman, M. Y. Mashor, H. Jaafar, R. A. A. Raof, and N. H. Harun, "Performance comparison between RGB and HSI linear stretching for tuberculosis bacilli detection in Ziehl-Neelsen tissue slide images," In 2009 IEEE International Conference on Signal and Image Processing Applications (ICSIPA),pp. 357-362,2009.
[40] D. Comaniciu and V. Ramesh, "Mean shift and optimal prediction for efficient object tracking," In Proceedings. 2000 International Conference on Image Processing, pp. 70-73 vol.3, 2000.
[41] A. Djouadi, O. Snorrason, and F. D. Garber, "The quality of training sample estimates of the Bhattacharyya coefficient," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 12, pp. 92-97, 1990.
[42] M. Castrill’on-Santana, O. D’eniz-Su’arez, C. G. Guerra, and M. Hernandez, "Real-time detection of multiple faces at different resolutions in video streams," Journal of Visual Communication and Image Representation, vol. 18, pp. 130-140, 2007.
[43] H. E. Lee, "Target tracking of autonomous mobile robots," 國立臺灣科技大學自動化及控制研究所碩士論文, 2010.