隨著科技的進步，網路、智慧型手機的發展，生物辨識技術的需求越來越大。而目前有許多生物辨識技術，像是人臉、指紋、虹膜、掌紋...等，其中許多已被廣泛使用。但與其他的識別方法相比，僅有人臉辨識能夠在不影響流程又不須接觸的情況下進行身分辨識。然而，在實際場景中，複雜的背景、人臉不同的拍攝角度、框選區域與大小等皆會影響到人臉辨識的辨識率，因此如何克服這些因素並且解決環境的干擾，以及提升人臉辨識的速度，降低辨識系統對使用者的影響就成為本論文的研究方向。
本論文分成了四個部分，分別為「人臉偵測」、「人臉正規化」、「人臉資料庫建立、訓練與辨識」、與「平行化加速」四個部分，主要目的是透過平行處理，建置一個擁有即時辨識能力，並且可以在非理想環境下實際運作的即時多標人臉辨識系統。
系統透過DVR(數位錄影監控系統)取得即時影像，在實際場景進行測試。運用本論文所提出之人臉辨識方法，人臉辨識率在正面時可達到93.41%，平均辨識率也達到85.80%，並可同時偵測多個人臉。在系統運作速度上，本論文亦進行程式效能瓶頸的改善，辨識速度從5fps提昇為13fps的運作速度，滿足了實際環境中即時處理的需求。

With all the advances in smart phone and internet technology, the demand for biometric recognition system is growing exponentially; there are many different identification technologies of biometric recognition system, such as face, fingerprint, palmprint. Many of them have been in widespread use. Compared with many other identification methods, only face recognition need not to make direct contact with an individual. However, there are many factors such as complex background, angles of human heads, selected area and size that will affect the recognition rate in non-ideal environments. Therefore how to reduce impacts of these factors to the recognition result and solve the interferences in real environments as well as to speed up the recognition process will be discussed in the thesis.
This research consists of four parts: “face detection”, “face normalization”, “establishing database of face, training, and recognition” and “parallel speedup” to build a real-time multiple face recognition system which could work under non-ideal environment. The main contribution of this work is focused on using OpenMP and Multi-Thread to speed up the face recognition processing.
This system captures real-time images from DVR (digital video recorder), and tests in real environments. Experimental results show that the system could capture multiple faces, achieves the overall average recognition rate of 85.80% and front-face recognition rate of 93.41%. The system is also improved on efficiency of recognition process by the proposed method. As a result, the face recognition processing frame rate increases from 5 fps to 13 fps.

[1] Kwok-Wai Wong, Kin-Man Lam and Wan-Chi Siu,”A Robust Scheme for Live Detection of Human Face in Color Images”, Signal Processing: Image Communication, The Netherlands, Vol. 18, No.2, pp.103-114, Feb. 2003
[2] D.Chai and A.Bouzerdoum, “A Bayesian Approach to Skin Color Classification in YCbCr Color Space,” TENCON 2000. Proceedings, IEEE, Kuala Lumpur Malaysia, Vol. 2, pp.421-424, Sept. 2000.
[3] P. Viola and M. J. Jones, "Robust Real-Time Face Detection," International Journal of Computer Vision, Vol. 57, pp. 137-154, May 2004.
[4] P. Viola and M. J. Jones, "Rapid Object Detection Using a Boosted Cascade of Simple Features," IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Vol. 1, pp. 511-518, 2001.
[5] P. Viola and M. J. Jones, "Fast and Robust Classification using Asymmetric AdaBoost and a Detector Cascade," NIPS, pp. 1311-1318, 2001.
[6] R. Lienhart, A. Kuranov and V. Pisarevsky, "Empirical Analysis of Detection Cascades of Boosted Classifiers for Rapid Object Detection", Technical report, MRL, Intel Labs, 2002.
[7] Y. Freund and R.E. Schapire, "A decision-theoretic generalization of on-line learning and an application to boosting," Journal of Computer and System Sciences, vol.55, pp. 119-139, 1997.
[8] R. E. Schapire, "The Strength of Weak learnability," Machine Learning, Vol. 5, pp. 197-227, 1990.
[9] V. Perlibakas, "Distance measures for PCA-based face recognition", Pattern Recognition Letters 25, 711 - 724, 2004.
[10] M. Kirby and L. Sirovich, "Application of the Karhunen-Loeve procedure for the characterization of human faces", IEEE Trans. Pat. Anal. Mach. Intell. 12, 1990.
[11] M.A. Turk and A.P. Pentland, "Face recognition using eigenfaces", Computer Vision and Pattern Recognition, 1991.
[12] T. F. Cootes, C, J, Taylor, D. H. Cooper, and J. Graham. “Active shape models-their training and application” Computer Vision and Image Understanding, 61(1):38-59, Jan. 1995.
[13] T.F. Cootes, D. Cooper, C.J. Taylor and J. Graham. “A Trainable Method of Parametric Shape Description” In Procs. British Machine Vision Conference, Springer, Verlag, pp.54-61, 1991.
[14] Tim Cootes. “An Introduction to Active Shape Models”. Chapter 7, Image Processing and Analysis, pp.223-248, Oxford University Press, 2000.
[15] T. Cootes, G. J. Edwards, and C. J. Taylor. “Active appearance models” In H.Burkhardt and B. Neumann, editors, 5th European Conference on Computer Vision, volume 2, pages 484–498. Springer, 1998.
[16] N. Otsu, "A threshold selection method from gray-level histogram," IEEE Transactions on System, Man, Cybernetics, 19(1):62-66, January 1978.
[17] 謝育書，「基於投影方法的人臉辨識之研究」，碩士論文，資訊工程所，國立清華大學，新竹(2008)。
[18] 呂元昊，「以SOPC為基礎之人臉辨識系統」，碩士論文，電子工程所，國立台灣科技大學，台北(2005)。
[19] 李易俊，「基於Gabor特徵及二維PCA之人臉辨識」，碩士論文，電腦與通訊工程所，國立成功大學，台南(2005)。.
[20] 劉翁昌，「複雜環境下之即時人臉偵測與辨識系統」，碩士論文，電子工程所，國立台灣科技大學，台北(2009)。
[21] 陳志銘，「利用環場及 PTZ攝影機建構室內環境監控系統作臉部辨識」，碩士論文，資訊工程所，國立中央大學，桃園(2008)。
[22] 張家豪，「以AAM與PCA為基礎之眼鏡特徵弱化方法於人臉辨識之改進」，碩士論文，資訊工程所，國立中央大學，桃園(2008)。
[23] 楊煒達，「簡易方法之少量人臉辨識系統」，碩士論文，資訊工程所，國立中央大學，桃園(2007)。
[24] 井上誠喜、八木伸行、林正樹、中須英府輔、三谷公二、奧井誠人，C語言數位影像處理，全華圖書，台北(2006)。