有鑒於實際的人臉偵測與辨識系統需要運作在真實複雜場景裡，環境中複雜的背景、人臉不同拍攝角度和人臉區域框選位置與大小皆會影響到人臉辨識率。因此本論文目的為克服這些因素並減少環境的干擾，建立能夠運作於真實環境的人臉辨識系統。在人臉偵測的部份結合了灰階人臉特徵偵測與彩色膚色密度計算，改善灰階特徵偵測的誤判情形。在人臉區域框選部份，利用人臉五官特徵定位，取出人臉區域並進行不同角度的旋轉校正，有效排除大部分複雜背景與降低人臉旋轉角度的影響。在資料庫訓練與人臉辨識的部份採用了主成份分析法(Principal Component Analysis, PCA)來降低人臉特徵資料量並利用人臉特徵空間投影係數的距離比對來決定辨識結果。

本論文結合了人臉偵測、人臉區域框選正規化與人臉辨識三個部份，實現複雜環境下之人臉辨識系統。系統在實際場景進行測試時，運用本論文所提出之人臉正規框選校正方法，人臉辨識率從原本的73.8%提昇至91.6%。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1.1fps提昇為9fps的運作速度，滿足了實際環境中即時處理的需求。

There is an urgent need for Face Detection and Recognition System (FDRS) to be used in complex situations. However, face recognition rates are often affected by complicated background, different head angles, variant face size and position. For building a Face Recognition System under real environment, this thesis aims to solve these problems and reduce the interferences occurred in the environment. The author combines gray-scale facial feature and skin color density to improve face detection yield. Also, at face identifiable region segmentation, the author uses facial feature to orientate facial region, and then modifies each with different angles. This method can efficiently exclude the dominant effects from the complex background information and the distortions via rotating angles. At database training and face recognition stage, the Principal Component Analysis method (PCA) is adopted to decompose and reduce the mass facial feature information. Moreover, the author uses the Euclidean distance of projection coefficients of facial eigen-space to determine the recognition results.

Overall, in this thesis, it combines face detection, face region normalization and face recognition to elaborate a face recognition system under complex scenarios. In real simulation, the proposed method improves face recognition rate from 73.8% up to 91.6%. Not only does the proposed method speed up the function efficiency from 1.1 fps up to 9fps but also fulfils the instant need in real settings.

[1]M. H. Yang and N. Ahuja, “Detecting Human Faces in Color Images,” IEEE Int’l Conf. Image Processing, pp. 127-139, Oct. 1998.
[2]R.L. Hsu, M. Abdel-Mottaleb, and AK. Jain, "Face Detection in Color Images," IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 24, no. 5, pp. 696-706, 2002.
[3]鄭凱方，「人臉可辨識度計算用於監控系統中人臉正面最佳影像判定」，碩士論文，資訊工程所，國立中央大學，桃園(2005)。
[4]K. Sobottka and I. Pitas, “Extraction of facial regions and features using color and shape information”, in Proc. 13th International Conference on Pattern Recognition, pp.421-425, Vienna, Austria, Aug. 1996.
[5]P. Viola and M. J. Jones, "Robust Real-Time Face Detection," International Journal of Computer Vision, Vol. 57, pp. 137-154, May 2004.
[6]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.
[7]P. Viola and M. J. Jones, "Fast and Robust Classification using Asymmetric AdaBoost and a Detector Cascade," NIPS, pp. 1311-1318, 2001.
[8]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.
[9]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.
[10]R. E. Schapire, "The Strength of Weak learnability," Machine Learning, Vol. 5, pp. 197-227, 1990.
[11]V. Perlibakas, "Distance measures for PCA-based face recognition", Pattern Recognition Letters 25, 711 - 724, 2004.
[12]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.
[13]M.A. Turk and A.P. Pentland, "Face recognition using eigenfaces," Computer Vision and Pattern Recognition, 1991.
[14]R. C. Gonzalez and R. E. Woods, “Digital Image Processing,” Second edition, Prentice Hall, 2002.
[15]N. Otsu, "A threshold selection method from gray-level histogram," IEEE Transactions on System, Man, Cybernetics, 19(1):62-66, January 1978.
[16]W. Niblack, “An Introduction to Digital Image Processing,” Prentice-Hall, Englewood Cliffs, New Jersey, pp. 115-116, 1986.
[17]J. Sauvola, M. Pietikainen, “Adaptive Document Image Binarization,” Pattern Recognition 33, p.225-236, 2000.
[18]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，中華民國95年。
[19]M. Sonka, V. Hlavac, and R. Boyle, Image Processing, Analysis, and Machine Vision, PWS Publishing , 1999.
[20]S. Jiazhen, S. E. Reichenbach, “Interpolation by asymmetric, two-dimensional cubic convolution,” IEEE International Conference on Image Processing, vol. 2, Sept. 2005.
[21]C. Garcia, and G. Tziritas. “Face Detection Using Quantized Skin Color Regions Merging and Wavelet Packet Analysis,” in IEEE Transactions on Multimedia, vol. 1. no. 3. September, pp. 264-277, 1999.
[22]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)。
[23]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)。
[24]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)。
[25]王志良、孟秀艷，人臉工程學，機械工業出版社(2008)。
[26]鐘國亮，影像處理與電腦視覺，東華書局(2002)
[27]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)。
[28]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)。
[29]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)。