在本論文中，我們提出了一個人臉辨識系統，當輸入兩張不同的人臉影像，根據人臉影像的區域二位元圖形以及同時使用兩種不同的公制學習法可以判斷所輸入的兩張影像是不是屬於同一個人，本篇系統是建立在影像中的人臉已經被偵測到，人臉的大小與位置都知道的情況下，將人臉單獨裁減下來輸入系統做驗證，同時人臉需要根據雙眼位置做幾何矯正，才能夠得到較好的結果。我們提出了一個新的公制學習法叫做「鑑別機率公制學習」，這個方法只考慮資料的類別外關係而不考慮資料類別內的關係，而且鑑別機率公制學習與巨大分界區最近鄰居公制學習法一起使用時，鑑別機率公制學習可以針對巨大分界區最近鄰居公制學習法對於偏離類別中心的資料表現較不好的情況進行補強，並且巨大分界區最近鄰居公制學習法也可以針對鑑別機率公制學習對於一般的人臉資料表現較不佳的情況補強而產生互補效用，同時我們也提出了另外一個新的人臉驗證的方法「K最近鄰居編碼」可以解決分界化K最近鄰居法的對於輸入兩張相同的影像一樣有可能判斷成不同人的情況，同時可以得到更好的效果。

In this thesis, we present a face verification system using local binary pattern and two difference metric learning method. With this system we can decide whether the two input face image is belong same person or not. This system is based on the position and size of face detected, and crops the facial part as the input of the verification system, the face image was normalized in geometric based on the location of two eyes. We present a new metric learning method called discriminant probability metric learning which only consider between relationships of data. When combine large margin nearest neighbor with discriminant probability metric learning, we can get better performance of outlier data than the method which only use large margin nearest neighbor and we also can get better performance of normal data than the method which only use discriminant probability metric learning. In this reason, we combined two metric learning methods to get better performance. We also present a new face verification method called K-nearest neighbor code which can solve the problem that when input same images, the verification result of marginalized K-Nearest Neighbor will be negative and can get better performance than previous verification methods.

[1] 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, No. 12, pp. 2037–2041, 2006.
[2] A. Bar-Hillel, T. Hertz, N. Shental, and D. Weinshall, “Learning a Mahanalobis metric from equivalence constraints,” Journal of Machine Learning Research, Vol. 6, pp. 937–965, 2005.
[3] A. B. Ashraf, S. Lucey and T. Chen, ”Learning Patch Correspondences for Improved Viewpoint Invariant Face Recognition,” IEEE International Conference on Computer Vision and Pattern Recognition, 2008
[4] C. Ben-Abdelkader and P. Griffin, “A Local Region-Based Approach to Gender Classification from Face Images,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 52-52, 2005.
[5] M. Belkin and P. Nyiogi, “Laplacian eigenmaps for dimensionality reduction and data representation,” Neural Comput, Vol. 15, No. 6, pp. 1373-1396, 2003.
[6] P. Belhumeur, J. Hespanha, and D. Kriegman, “Eigenfaces vs. Fisherfaces: Recognition Using Class Sspecific Linear Projection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 19, No. 7, pp. 711–720, 1997.
[7] D. Cai, X. He, K. Zhou, J.Han and H. Bao, “Locality Sensitive Discriminant Analysis,” International Joint Conference on Artificial Intelligence, 2007
[8] C. Cortes and V. Vapnik, “Support-Vector Networks,” Machine Learning, Vol. 20, 1995.
[9] P. Comon, ”Independent Component Analysis: a new concept?,” IEEE Transactions on Signal Processing, Vol. 36, No. 3, pp. 287-314, 1994.
[10] Z. Cao, Q. Yin, X. Tang and J. Sun, “Face Recognition with Learning-based Descriptor,” IEEE Conference Computer Vision and Pattern Recognition, 2010
[11] T. M. Cover and P. E. Hart, “Nearest neighbor pattern classification,” IEEE Transactions on Information Theory , Vol. 13, No. 1, pp. 21–27, 1967
[12] D. Donoho and C. Grimes, “Hessian Eigenmaps: Locally Linear Lmbedding Techniques for High-Dimensional Data,” Proc Natl Acad Sci, Vol. 100, No. 10, pp5591-5596, 2003
[13] J. Davis, B. Kulis, P. Jain, S. Sra, and I. Dhillon, ”Information Theoretic Metric Learning,” International Conference on Machine Learning, 2007
[14] N. Dalal and B. Triggs, “Histograms of Oriented Gradients for Human Detection,” IEEE Conference Computer Vision and Pattern Recognition, vol. 2, pp. 886-893, 2005
[15] A. Frome, Y. Singer, F. Sha, and J. Malik, “Learning globally-consistent local distance functions for shape-based image retrieval and classification,” International Conference on Computer Vision, pp. 1-8, 2007.
[16] B. Fröba and A. Ernst, “Face Detection with the Modified Census Transform,” IEEE Conference on Face and Gesture Recognition, pp. 91-96, 2004
[17] A. Globerson and S. Roweis, “Metric learning by collapsing classes,” Neural Information Processing Systems, 2006
[18] M. Guillaumin, J. Verbeek and C. Schmid, “Is That You? Metric Learning Approaches for Face Identification,” IEEE International Conference on Computer Vision, 2009.
[19] D. G. Lowe, “Distinctive Image Features from Scale-Invariant Keypoints,” International Journal of Computer Vision, pp. 91-110, 2004.
[20] M. Jones and P. Viola, “Fast Multi-View Face Detection,” Tech. Rep., 2003.
[21] P. Jain, B. Kulis, I. S. Dhillon, and K. Grauman, “Online Metric Learning and Fast Similarity Search,” Neural Information Processing Systems, 2008
[22] N. Kumar, A. C. Berg, P. N. Belhumeur and S. K. Nayar, “Attribute and Simile Classifiers for Face Verification,” IEEE International Conference on Computer Vision, 2009.
[23] A. Li, S. Shan, X. Chen and W. Gao, “Maximizing Intra-individual Correlations for Face Recognition Across Pose Differences,” IEEE Conference Computer Vision and Pattern Recognition, 2009.
[24] K. Levi and Y. Weiss, “Learning Object Detection from a Small Number of Examples: The Importance of Good Features,” IEEE Conference Computer Vision and Pattern Recognition, pp. 53–60, 2004.
[25] P. Lee, G. Hsu and Y. Hung, “Face Verification and Identification using Facial Trait Code,” pp.1613~1620, IEEE Conference Computer Vision and Pattern Recognition, 2008.
[26] S. T. Roweis and L. K. Saul, ”Nonlinear Dimensionality Reduction by Locally Linear Embedding,” Science, Vol. 290, pp.2323-2326, 2000.
[27] J. Shi and C. Tomasi, “Good Features to Track,” IEEE Conference Computer Vision and Pattern Recognition, 1994
[28] M. Schultz and T. Joachims, ” Learning a distance metric from relative comparisons,” Neural Information Processing Systems, 2003.
[29] S. Shalev-Shwartz, Y. Singer, and A. Y. Ng, “Online and batch learning of pseudo-metrics,” International Conference on Machine Learning, 2004.
[30] J. B. Tenenbaum, V. de Silva and J. C. Langford, “A Global Geometric Framework for Nonlinear Dimensionality Reduction,” Science, Vol. 290, pp. 2319–2323. 2000
[31] K. Q. Weinberger, J. Blitzer and L. K. Saul, “Distance Metric Learning for Large Margin Nearest Neighbor Classification,” Journal of Machine Learning Research, Vol. 10, pp. 209-244, 2009.
[32] R. Wang and X. Chen, “Manifold Discriminant Analysis,” IEEE Conference Computer Vision and Pattern Recognition, Vol. 0, pp. 429-436, 2009
[33] B. Yao and S. Lao, “Logit-RankBoost with Pruning for Face Recognition,” IEEE International Conference on Automatic Face and Gesture Recognition, 2008
[34] S. Yan, S. Shan, X. Chen, and W. Gao, “Locally Assembled Binary (LAB) Feature with Feature-Centric Cascade for Fast and Accurate Face Detection,” IEEE Conference Computer Vision and Pattern Recognition, 2010, pp. 1-7, 2008.
[35] S. Yan, D. Xu, B. Zhang, and H. Zhang, “Graph Embedding: A General Framework for Dimensionality Reduction,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 29, No. 1, pp. 40-41, 2007.