主動形狀模型(Active Shape Models，簡稱ASMs)是一種描繪物件形狀的統計模型。ASMs中包含了一個平均形狀及一組形狀變異模式(shape variation modes)，且利用平均形狀與形狀變異模式之線性組合，在疊代過程中逐次變化模型，使其能與目標影像相符。由於最佳模型參數可精確地合成出目標物件之外形，因此如何搜尋並找出測試影像之最佳模型參數，為辨識相關應用之前的重要步驟。
在本論文中，我們提出一個新的主動形狀模型匹配演算法(ASMs fitting algorithm)，此演算法包含了三部分。第一部分提供初始模板的選擇，且此部分可使ASMs由較佳的初始形狀開始搜尋。第二部分則以計算相關度的方式，提出新的目標函數(objective function)，以判斷系統是否收斂。在第三部分，我們在搜尋之前先訓練一參照表(reference table)，並且使用此參照表於搜尋階段更新模型參數。相較於原始ASMs，我們的演算法降低了23.60%的形狀誤差，增加41.38%(約1.1秒)的搜尋時間。與原始AAMs相比，我們的演算法降低了18.76% 的形狀誤差，並且也減少了59.58% (約5.4秒)的搜尋時間。而與robust AAMs相比，我們的演算法降低了16.98% of形狀誤差，且減少90.98% (約37.3秒)的搜尋時間。

Active shape models (ASMs) are the statistical models for representing the shapes of objects. ASMs include a mean shape and a set of shape variation modes, and iteratively deform to fit the target image with linear combination of the mean shape and the variation modes. Because the optimal model parameters can exactly synthesize the shapes of target objects, how to search and find out the optimal model parameters of test images is an important step before recognition.
In this paper, we propose a novel ASMs fitting algorithm which includes three parts. In the first part, it provides initial modes selection, and it could make the ASMs search from better initial shapes. The second part proposes a novel objective function where the convergence of the system is determined based on the correlation coefficient. The last part, we train a reference table before search, and update the model parameters with the reference table during the search process. Compare with the original ASMs, our algorithm reduces 23.60% of the shape error, and increases 41.38% (about 1.1s) of search time. Compare with the original AAMs, our algorithm reduces18.76% of the shape error, and reduces 59.58% (about 5.4s) of search time. Compare with the robust AAMs, our algorithm reduces 16.98% of the shape error, and reduces 90.98% (about 37.3s) of search time.

參考文獻
[1] T. F. Cootes and C. J. Taylor, ”Active shape models – smart snakes,” in Proc. British Machine Vision Conference, pp. 266-275, 1992.
[2] T. F. Cootes , C. J. Taylor, D. H. Cooper, and J. Graham, “Training models of shape from sets of examples,” in Proc. British Machine Vision Conference, pp. 9-18, 1992.
[3] T. F. Cootes, C. J. Taylor, D. H. cooper, and J. Graham, “Active shape models – their training and application,” Computer Vision and Image Understanding, vol. 61, no. 1, pp. 38-59, January 1995.
[4] W. Wang, B. Yin, T. Hu, and K. Wang, “Global texture constrained active shape models for image interpretation,” in Proc. Int’l Conf. Neural Networks and Signal Processing, vol. 2, pp. 1031-1034, 2003.
[5] S. Yan, C. Liu, S. Z. Li, H. Zhang, H.-Y. Shum, and Q. Cheng, “Face alignment using texture-constrained active shape models,” Image and Vision Computing, vol. 21, no. 1, pp. 69-75, January 2003.
[6] M. Wimmer, F. Stulp, S. Pietzsch, and B. Radig, “Learning local objective function for robust face model fitting,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 30, no. 8, pp. 1357-1370, 2008.
[7] J. Liu and J. K. Udupa, “Oriented active shape models,” IEEE Trans. Medical Imaging, vol. 28, no. 4, pp. 571-584, 2009.
[8] T.F. Cootes, G. J. Edwards, and C. J. Taylor, "Comparing active shape models with active appearance models", in Proc. British Machine Vision Conference, vol. 1, pp. 173-182, 1999.
[9] T. F. Cootes, G. J. Edwards, and C. J. Taylor, “Active appearance models,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 23, no. 6, pp. 681-685, 2001.
[10] Y.-F. Lin and C.-W. Tang, “A robust active appearance models search algorithm,” in Proc. Advances in Multimedia Information Processing - PCM, vol. 5353, pp. 879-882, December 2008.
[11] X. Gao, Y. Su, X. Li, and D. Tao, “A review of active appearance models,” IEEE Trans. Systems, Man, and Cybernetics, Part C: Applications and Reviews, vol. 40, no. 2, pp. 145-158, 2010.
[12] R. Donner, M. Reiter, G. langs, P. Peloschek, and G. Bischof, “Fast active appearance model search using canonical correlation analysis,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 10, pp. 1690-1694, October 2006.
[13] X. W. Hou, S. Z. Li, H. J. Zhang, and Q. S. Cheng, “Direct appearance models,” in Proc. Computer Vision and Pattern Recognition, vol. 1, pp. I-828-I-833, 2001.
[14] J. Sung, T. Kanade, and D. Kim, “A unified gradient-based approach for combining ASM into AAM,” Int’l J. Computer Vision, vol. 75, no. 2, pp. 297-309, 2007.
[15] M. Kass, A. Witkin, and D. Terzopoulos, “Snakes: active contour models,” Int’l J. of Computer Vision, vol. 1, no. 4, pp. 321-331, Jan. 1988.
[16] M. Wimmer, S. Fujie, F. Stulp, T. Kobayashi, and B. Radig, “An ASM fitting method based on machine learning that provides a robust parameter initialization for AAM fitting,” in Proc. IEEE Int’l Conf. Automatic Face and Gesture Recognition, pp. 1-6, Apr. 2009.
[17] Q. D. Hguyen and M. Milgram, “Multi features active shape models for lip contours detection,” in Proc. Int’l Conf. Wavelet Analysis and Pattern Recognition, vol. 1, pp. 172-176, Aug. 2008.
[18] X. Pan, J. Kong, and A. W.-C. Liew, “Lip contour extraction based on support vector machine,” in Proc. Image and Signal Processing, vol. 3, pp. 697-700, July 2008.
[19] S. Stillittano, V. Girondel, and A. Caplier, “Inner and outer lip contour tracking using cubic curve parametric models,” in Proc. Int’l Conf. Image Processing, pp. 2469-2472, Feb. 2010.
[20] L. Yaling and D. Minghui, “Lip contour extraction based on manifold,” in Proc. Int’l Conf. MultiMedia and Information Technology, pp. 229-232, June 2009.
[21] W. C. Ooi, C. Jeon, K. Kim, D. K. Han, and H. Ko, “Effective lip localization and tracking for achieving multimodal speech recognition,” in Proc. IEEE Int’l Conf. Multisensor Fusion and Integration for Intelligent systems, pp. 90-93, 2008.
[22] T. F. Cootes, G. J. Edwards, and C. J. Taylor, “Active appearance models,” in Proc. European Conf. computer Vision, pp. 484-498, 1998.
[23] I. Mattthews and S. Baker, “Active appearance models revisited,” Int’l J. Computer Vision, vol. 60, no. 2, pp. 135-164, 2004.
[24] T. F. Cootes and P. Kittipanya-ngam, “Comparing variations on the active appearance model algorithm,” in Proc. British Machine Vision Conference, vol. 2, pp. 837-846, 2002.
[25] T. Kanade, J. Cohn, and Y. Tian, “Comprehensive database for facial expression analysis,” in Proc. IEEE Int’l Conf. on Automatic Face and Gesture Recognition, pp. 46-53, 2000.
[26] T. Kanade and J. F. Cohn, Automatic Face Analysis, http://www-2.cs.cmu.edu./~face/, June 2010.