本論文提出了一種透過利用人臉地標點及深度轉移(Depth Warping)的方法來解決跨角度人臉辨識的問題。不同於多數基於3D的解決方法，本論文著重於利用人臉地標點所提供的資訊實現深度轉移來取代通常需要昂貴計算成本的三維人臉重建方法。給定一張正面人臉影像經過地標點偵測，與三維人臉模型資料庫內的人臉地標點組相比較，來尋找資料庫內的相近臉型，因此正面人臉影像可以透過相近臉型的人臉模型應用深度轉移來重建出一個人臉模型。對齊有角度的測試人臉影像時，利用測試人臉影像上的地標點，使重建的人臉模型旋轉對齊相同角度，透過同時考慮Yaw、Pitch、Roll等三個角度來調整人臉模型旋轉角度，可使人臉模型準確的與測試影像上人臉對齊。為了處理眼鏡的問題本論文透過雙層樹狀結構模型(B-TSM)來偵測人臉上的眼鏡。本論文所提出的方法經標準人臉資料庫之測試，含PIE與Multi-PIE，並與近幾年方法比較確認其有效性。

An approach exploiting facial landmarks and depth warping is proposed for cross-pose recognition. Different from most 3D based cross-pose recognition, the proposed one uses facial landmarks extensively and replaces the computationally expensive 3D reconstruction by depth warping. Given a face in the gallery with landmarks detected, the spatial distribution of landmarks is used to find a similar face in a 3D face database so that the gallery face can be approximately reconstructed using depth warping with the most similar 3D face. When matching to a probe, the depth-warped faces are rotated to align to the same orientation revealed by the landmarks on the probe. The orientation is justified by yaw, pitch and roll of the probe, enabling an accurate alignment to the probe face. To handle the cases with eyeglasses, we propose a Bilayer Tree Structured Model (B-TSM) for detecting faces with eyeglasses. The proposed method is validated on PIE and Multi-PIE, and compared to contemporary approaches to show its efficacy.

[1] X. Zhang and Y. Gao, “Face recognition across pose: A review,” PR, vol. 42, pp. 2876–2896, Nov. 2009.
[2] G.S. Hsu, H.C. Peng, K.H. Chang, “Landmark Based Facial Component Reconstruction for Recognition across Pose,” CVPRW, 2014.
[3] X. Zhu, D. Ramanan. “Face detection, pose estimation and landmark localization in the wild,” CVPR Providence, Rhode Island, June 2012.
[4] P. Felzenszwalb, D. McAllester, D. Ramanan, “A discriminatively trained, multiscale, deformable part model,” in CVPR 2008.
[5] S. Ioffe, D. Forsyth, “Mixtures of trees for object recognition,” CVPR, 2001.
[6] A. Li, S. Shan, W. Gao, “Coupled Bias–Variance Tradeoff for Cross-Pose Face Recognition,” in TIP , vol.21, no.1, pp.305-315, Jan. 2012
[7] A. Asthana, T.K. Marks, M.J. Jones, K.H. Tieu, M. Rohith, “Fully automatic pose-invariant face recognition via 3D pose normalization,” in ICCV , vol., no., pp.937-944, 6-13 Nov. 2011
[8] V. Blanz and T. Vetter. “A morphable model for the synthesis of 3D faces,” In SIGGRAPH. 1999. 3,4,6,7
[9] W. Zhang, S. Shan, W. Gao, X. Chen, and H. Zhang. “Local gabor binary pattern histogram sequence (LGBPHS): A novel non-statistical model for face representation and recognition.” TEEE TCCV, 2005
[10] A. Moeini, H. Moeini, “Real-World and Rapid Face Recognition Toward Pose and Expression Variations via Feature Library Matrix,” in TIFS, vol.10, no.5, pp.969-984, May 2015
[11] C. Ding, C. Xu, D. Tao, “Multi-Task Pose-Invariant Face Recognition,” in TIP , vol.24, no.3, pp.980-993, March 2015
[12] U. Prabhu, J. Heo, M. Savvides, “Unconstrained Pose-Invariant Face Recognition Using 3D Generic Elastic Models,” in TPAMI, vol.33, no.10, pp.1952-1961, Oct. 2011
[13] J. Heo, M. Savvides, “Gender and Ethnicity Specific Generic Elastic Models from a Single 2D Image for Novel 2D Pose Face Synthesis and Recognition,” in TPAMI, vol.34, no.12, pp.2341-2350, Dec. 2012
[14] R. Abiantun, U. Prabhu, M. Savvides, “Sparse Feature Extraction for Pose-Tolerant Face Recognition,” in TPAMI, vol.36, no.10, pp.2061-2073, Oct. 2014
[15] P. Phillips, P. J. Flynn, T. Scruggs, K. W. Bowyer, J. Chang, K. Hoffman, J. Marques, J. Min, and W. Worek, “Overview of the face recognition grand challenge,” in CVPR, vol. 1, 2005, pp. 947–954.
[16] X. Tan and B. Triggs, “Enhanced local texture feature sets for face recognition under difficult lighting conditions,” in TIP, vol. 19, p. 1635–1650, June 2010.
[17] J. Wright, A. Y. Yang, A. Ganesh, S. S. Sastry, and Y. Ma, “Robust face recognition via sparse representation,” in TPAMI, vol. 31, pp. 210–227, 2009.
[18] C. Xie, M. Savvides, and B. V. K. V. Kumar, “Redundant classdependence feature analysis based on correlation filters using frgc 2.0 data,” in Proc. CVPR. Workshops, Jun. 2005, vol. 3, pp. 153–158.
[19] A.M. Ali, “A 3D-Based Pose Invariant Face Recognition at a Distance Framework,” in TIFS , vol.9, no.12, pp.2158-2169, Dec. 2014
[20] E. Mostafa and A. A. Farag, “Complex Bingham distribution for facial feature detection,” in Computer Vision (Lecture Notes in Computer Science), vol. 7584. Berlin, Germany: Springer-Verlag, 2012, pp. 330–339.
[21] A. Asthana, M. J. Jones, T. K. Marks, K. H. Tieu, and R. Goecke, “Pose normalization via learned 2D warping for fully automatic face recognition,” in Proc. BMVC, 2011, pp. 1–11.
[22] F. Juefei-Xu, K. Luu, M. Savvides, “Spartans: Single-Sample Periocular-Based Alignment-Robust Recognition Technique Applied to Non-Frontal Scenarios,” in TIP , vol.24, no.12, pp.4780-4795, Dec. 2015
[23] R. Gross, I. Matthews, S. Baker, “Appearance-based face recognition and light-fields,” in TPAMI, vol.26, no.4, pp.449-465, April 2004
[24] S.J.D. Prince, J. Warrell, J.H. Elder, F.M. Felisberti, “Tied Factor Analysis for Face Recognition across Large Pose Differences,” in TPAMI , vol.30, no.6, pp.970-984, June 2008
[25] A. Li, S. Shan, X. Chen, W. Gao, “Cross-pose face recognition based on partial least squares”, Pattern Recognition Letters, Volume 32, Issue 15, 1 November 2011, Pages 1948-1955
[26] A. Asthana, S. Zafeiriou, S. Cheng and M. Pantic. “Robust Discriminative Response Map Fitting with Constrained Local Models.” in CVPR, 2013.
[27] A. Telea, “An Image Inpainting Technique Based on the Fast Marching Method.” Journal of Graphics, GPU, and Game Tools 9 1, pp 23-34, 2004.
[28] R. Gross, I. Matthews, J. Cohn, T. Kanade, and S. Baker, “Multi-pie,” Proc. in AFGR, pp. 1,8,17–19, Sept. 2008.
[29] T. Sim, S. Baker, and M. Bsat, “The CMU pose, illumination, and expression database,” in AFGR, 2002, pp. 46–51
[30] G.S. Hsu; K.H. Chang; S.C. Huang; S.L. Chung, “Face detection and landmark localization using Bilayer Tree Structured Model,” in ICIP, 2015.
[31] A. Sharma, M. A. Haj, J. Choi, L. S. Davis, and D. W. Jacobs, “Robust pose invariant face recognition using coupled latent space discriminant analysis,” Comput. Vis. Image Understand., vol. 116, no. 11, pp. 1095–1110, Nov. 2012.
[32] D. González-Jiménez and J. L. Alba-Castro, “Toward pose-invariant 2-D face recognition through point distribution models and facial symmetry,” TIFS, vol. 2, no. 3, pp. 413–429, Sep. 2007.
[33] H. T. Ho and R. Chellappa, “Pose-invariant face recognition using Markov random fields,” IEEE Trans. Image Process., vol. 22, no. 4, pp. 1573–1584, Apr. 2013.
[34] X. Zhang and Y. Gao, “Heterogeneous specular and diffuse 3-D surface approximation for face recognition across pose,” IEEE Trans. Inf.
Forensics Security, vol. 7, no. 2, pp. 506–517, Apr. 2012.