本論文將以實際應用的人臉辨識系統為考量，進行相關的研究探討，主要以探討變化光源環境下的臉部辨識系統之設計與製作，因無法得知使用者會在何種光源環境下進行臉部辨識，本研究將利用擁有大量不同光源之標準人臉資料庫進行多種可能情況的模擬，並提出有效的解決方案。全域Log-DCT已被證明可有效在不同光源下進行人臉辨識，本論文針對光源變化和人臉定位偏差的問題透過局部區塊特徵的結合進行人臉辨識。本研究的目標包含以下：(1)將全域Log-DCT與局部區塊的結合效能進行定義，(2)檢視人臉偵測系統的人臉定位偏差對於人臉辨識的衝擊。令人滿意的實驗結果顯示於CMU PIE人臉資料庫由於近乎完美的臉部定位，這表示全域Log-DCT與局部區塊的結合當定位無偏差時可有效解決光源變化的影響，然而在測試FRGC 2.0資料庫時，卻降低了此組合的評價，因FRGC 2.0提供了定位的偏差、角度和表情變化。這反映出實際的人臉辨識系統無法不考慮這些參數，局部區塊的搜尋配對針對以上問題進行修正並被證明有效的提升人臉辨識於此論文的研究。

Based on the holistic Log-DCT features, which are proven effective for face recognition across illumination conditions, this research considers the same features combined with local square patches for face recognition across illumination and with imprecise face localization. The objectives of this research include the following: (1) define the performance upper bound attainable by the combination of holistic Log-DCT features and local patches, (2) investigate the impacts on the performance from imprecise face localization contributed by a face detector. Satisfactory results are shown from the experiments on the CMU PIE database which offers faces with almost perfect localization, revealing that the combination of Log-DCT features and local patches can be an effective solution for recognizing perfectly localized faces across illumination. However, the performance degrades substantially when evaluating the combination on the FRGC 2.0 database, which offer faces with imprecise localization and variations on pose and expression, reflecting the fact that an actual face recognition system cannot leave alone these parameters. A local alignment and masking scheme is proposed to tackle the problems caused by these parameters, and is proven effective in an extensive experimental study.

[1]et al, B. H. Face recognition: component-based versus global approaches Computer Vision and Image Understanding, 91:6-21, 2003
[2]et al, M. B. Face recognition by independent component analysis IEEE Trans. on Neural Networks, 13(6):1450–1454, 2002
[3]et al, T. K. Component-based LDA Face Description for Image Retrieval and MPEG-7 Standardisation Image and Vision Computing, 23(7):631-642, 2005
[4]A. Abbas, M. I. Khalil, S. A. H. M. A. F. ILLUMINATION INVARIANT FACE RECOGNITION IN LOGARITHM DISCRETE COSINE TRANSFORM DOMAIN ICIP, 2009
[5]A. Pentland, B. Moghaddam, T. S. & Turk, M. View based and modular eigenspaces for face recognition Proceedings of IEEE CVPR, pp. 84-91, 1994
[6]A. S. Georghiades, P. N. Belhumeur & Jacobs, D. W. From few to many: illumination cone models for face recognition under variable lighting and pose IEEE Trans. Pattern Anal. Mach. Intel., vol. 23, no. 6, pp. 630–660, Jun., 2001
[7]Basri, R. & Jacobs, D. W. Lambertian reflectance and linear subspaces IEEE Trans. Pattern Anal. Mach. Intell., vol. 25, no. 2, pp.218–233, Feb., 2003
[8]Belhumeur, P. N. & Kriegman, D. J. What is the set of images of an object under all possible illumination conditions Int. J. Comput. Vis., vol. 28, no. 3, pp. 245–260, Jul., 1998
[9]Fabrizia M. de S. Matos, L. V. B. & v. d. Poel, J. Face Recognition Using DCT Coefficients Selection ACM 978-1-59593-753-7/08/0003, 2008
[10]Gonzalez, R. C. & E.Woods, R. Digital Image Processing Reading, 1992
[11]H.K. Ekenel, A. P. Video-based Face Recognition Evaluation in the CHIL Project –Run 1 7th Intl. Conf. on Automatic Face and Gesture Recognition (FG 2006), Southampton, UK, 2006
[12]H.K. Ekenel, R. S. A Generic Face Representation Approach for Local Appearance based Face Verification CVPR IEEE Workshop on FRGC Experiments, 2005[13]H.K. Ekenel, R. S. Local Appearance based Face Recognition Using Discrete Cosine Transform EUSIPCO 2005, Antalya, Turkey, 2005
[14]Hafed, Z. M. & Levine, M. D. Face recognition using the discrete cosine transform International Journal of Computer Vision, 43(3), 2001
[15]HAFED, Z. M. & LEVINE, M. D. Face Recognition Using the Discrete Cosine Transform International Journal of Computer Vision 43(3), 167–188, 2001
[16]HazIm Kemal Ekenel, Mika Fischer, R. S. Face Recognition in Smart Rooms ICMI-MLMI, 2007
[17]HazIm Kemal Ekenel, R. S. Analysis of Local Appearance-based Face Recognition: Effects of Feature Selection and Feature Normalization CVPR, 2006[18]Horn, B. K. P. Robot Vision Cambridge, 1986
[19]J.H.P.N. Belhumeur, D. K. Eigenfaces vs. fisherfaces: Recognition using class specific linear projection IEEE Trans. on PAMI, 19(7):711–720, 1997
[20]Land, E. H. & McCann, J. J. Lightness and retinex theory J. Opt. Soc. Amer., 1971
[21]M. Turk, A. P. Eigenfaces for recognition Journal of Cognitive Science, pages 71–86, 1991
[22]Nayar, S. K. & Bolle, R. M. Reflectance based object recognition Int. J. Comput. Vis., 1996
[23]Nefian, A. A Hidden Markov Model-based Approach for Face Detection and Recognition PhD thesis, Georgia Institute of Technology, 1999
[24]P. N. Belhumeur, J. P. Hespanha & Kriegman, D. J. Eigenfaces versus Fisherfaces: recognition using class specific linear projection IEEE Trans. Pattern Anal. Mach. Intell., vol. 19, no. 7, pp. 711–720, Jul., 1997
[25]Pan, Z. & Bolouri, H. High speed face recognition based on discrete cosine transforms and neural networks Technical report, University of Hertfordshire, UK, 1999
[26]Pizer, S. M. & Amburn, E. P. Adaptive histogram equalization and its variations Comput. Vis. Graph., Image Process., vol. 39, no. 3, pp. 355–368, 1987
[27]R. Brunelli, T. P. Face Recognition: Features versus Templates IEEE Trans. on PAMI, 13(10): 1042-1052, 1993
[28]R. Gottumukkal, V. A. An improved face recognition technique based on modular PCA approach Pattern Recognition Letters, 25(4), 2004
[29]S. Shan, W. Gao, B. C. & Zhao, D. Illumination normalization for robust face recognition against varying lighting conditions in Proc. IEEE Workshop on AMFG, 2003, pp. 157–164., 2003
[30]Sanderson, C. & Paliwal, K. K. Features for robust facebased identity verification Signal Processing, 83(5), 2003
[31]Savvides, M. & Kumar, V. Illumination normalization using logarithm transforms for face authentication in Proc. IAPR AVBPA, 2003, pp. 549–556, 2003
[32]Scott, W. L. Block-level Discrete Cosine Transform Coefficients for Autonomic Face Recognition PhD thesis, Louisiana State University, USA, 2003
[33]Weilong Chen, M. J. E. & Wu, S. Illumination Compensation and Normalization for Robust Face Recognition Using Discrete Cosine Transform in Logarithm Domain IEEE, 2006
[34]Xie, X. & L, K.-M. Face recognition under varying illumination based on a 2D face shape model Pattern Recognit., to be published
[35]Y. Adini, Y. Moses & Ullman, S. Face recognition: the problem of compensating for changes in illumination direction IEEE Trans. Pattern Anal. Mach. Intell., vol. 19, no. 7, pp. 721–732, Jul., 1997
[36]Yuehui Chen, S. J. & Abraham, A. Face Recognition Using DCT and Hybrid Flexible Neural Tree IEEE, 2005
[37]Zhang, L. & Samaras, D. Face recognition under variable lighting using harmonic image exemplars in Proc. IEEE Conf. Computer Vision and Pattern Recognition, vol. 1, 2003, pp. 19–25., 2003