隨著科技的日新月異之下與數位影像的普及以及數位影像處理技術的突飛猛進，指紋辨識、人臉辨識、虹膜辨識等相關的生物認證研究也一一的出現，在這之中人臉辨識可以是最便利的認證，因為不需要有任何的碰觸，只需要透過攝影機將影像抓取進來就可以得到辨識所需要的資料，因此可大量的應用於機器人、監視器、攝影機、平板電腦與手機等產品上。
本篇論文為整合Adaboost人臉偵測與DCT(Discrete Cosine Transform)人臉辨識即時人臉偵測與辨識系統，一開始先透過色彩空間轉換將原本彩色影像轉換為灰階影像，偵測部分是以Opencv的Adaboost人臉偵測為結果，進入辨識演算法後會先執行縮圖步驟，這個步驟可以使執行速度變快並且統一全部圖片大小，接下來背景消除是因為在人臉偵測時所框選的範圍包含了很多不是人臉的部分，所以會使用大範圍去除與三角型填黑的方式去除，演算法核心是利用全影像做DCT轉換，利用DCT轉換把空間域轉換成頻率域係數，並保存每一個係數再透過這些係數去做比對，進而產生出辨識的結果。由於光線強弱對於辨識的影響很大，加入了亮度與對比的調整，來做為光線補償的方式，另外在資料庫分析上，我們會去計算頻率域係數的histogram分佈以取得所要比對的座標點位置，這樣一來就不用每一個像素點都去做比對，以增加整體運算速度。
在實際辨識結果方面，我們使用標準人臉資料庫與自製人臉資料庫以及動態影像來做測試，結果顯示辨識率都有90%以上的效果，在標準資料庫可達到95%的辨識率，本論文所提出的演算法以少量得特徵點個數就達到理想的辨識率，應用在平板時由於比電腦測試時的環境更加複雜，並且硬體效能較差的緣故，使得平板在辨識率與速度都有較差的結果。

This paper considers the joint usage of Adaboost and the Discrete Cosine Transform (DCT) for a real-time human face detection and recognition system. Firstly, the original colored images were transformed into gray-level images through a color space transformation. Subsequently, images obtained by using Adaboost from the Open Source Computer Vision library (OpenCV) for human face detection were reduced in size while the face recognition algorithm was in progress, which resulted in a speed up in operation and making the sizes of all images consistent. Further, the background from the images were extracted by removing a large background area or filling the triangle areas with solid colours, as a handful of non-human face parts were included in the selected areas at human face detection. Afterwards, the kernel DCT algorithm was applied to the entire image where coefficients from the spatial domain were transformed to the frequency domain, and all coefficients were stored and taken into account while matching to generate the result of human face recognition. As brightness had a significant influence on recognition, the adjustment of brightness and contrast was implemented as a way of light compensation. In data analysis, we proposed to calculate the histogram distribution of the frequency domain coefficients to locate the coordinate points for matching.
As for the result of recognition in practice, we used the standard human face data base, the self-made human face data base, and dynamic images for testings. All tests showed an over 90% recognition rate, and the recognition rate in the standard human face data base was even over 95%. The algorithm proposed in this paper could achieve an ideal rate of recognition by using a small number of features.

中文摘要 i
ABSTRACT ii
誌謝 iii
目錄 iv
表目錄 vi
圖目錄 vii
一、 緒論 1
1.1 前言 1
1.2 研究動機與目的 2
1.3 論文架構 3
二、相關研究與探討 4
2.1人臉偵測相關研究 4
2.2人臉辨識相關研究 6
三、 人臉偵測 10
3.1 Adaboost人臉偵測演算法 10
3.1.1矩形特徵 11
3.1.2積分影像 12
3.1.3 AdaBoost 14
3.1.4 層疊分類器 16
3.2特徵點搜尋之人臉偵測演算法 18
3.2.1搜尋特徵點 18
A.眼睛交界點 20
B.眉毛過濾特徵點 21
C.膚色偵測過濾特徵點 23
D.邊緣偵測過濾特徵點 26
3.2.2人臉的框選 28
3.3人臉偵測演算法比較 32
四、提出人臉辨識演算法 33
4.1人臉資料庫建立 33
4.1.1 色彩空間轉換 34
4.1.2縮圖 35
4.1.3去除背景 41
4.1.5離散餘弦轉換 48
4.1.6資料儲存 50
4.2人臉辨識 51
4.2.1分析資料庫的histogram分佈 52
4.2.2調整亮度與對比 55
4.2.3計算差異度 57
五、人臉辨識結果與比較 60
5.1 靜態影像 60
5.2 動態影像 67
5.3 平板的應用 74
六、結論 79
參考文獻 80

[1] Jizeng Wang and Hongmei Yang,” Face Detection Based on Template Matching and 2DPCA Algorithm,” Image and Signal Processing, 2008. CISP ''08. Congress,pp. 575 – 579,2008

[2] Paul Viola and Michael Jones,” Rapid Object Detection using a Boosted Cascade of Simple Features,” Computer Vision and Pattern Recognition, 2001. CVPR 2001. Proceedings of the 2001 IEEE Computer Society Conference ,pp. I-511 - I-518,2001

[3] Peter Lim, Liyanage C. De Silva, Liu Jing, and Li Li Ling,” Multimodal Approach to Human-Face Detection and Tracking,” IEEE Transactions on Industrial Electronics, vol. 55, no. 3,pp.1385-1393, Mar. 2008

[4] David Zhang, Hui Peng, Jie Zhou, and Sankar K. Pal,” A Novel Face Recognition System Using Hybrid Neural and Dual Eigenspaces Methods,” IEEE Transactions on Systems, Man, And Cybernetics—Part A: Systems And Humans, vol. 32,no. 6,pp.787-793, Nov. 2002

[5] Giancarlo Iannizzotto and Francesco La Rosa,” Competitive Combination of Multiple Eye Detection and Tracking Techniques,” IEEE Transactions on Industrial Electronics, vol. 58,no. 8, pp.3151-3159,Aug.2011


[6] Keun-Chang Kwak and Witold Pedrycz,” Face Recognition Using an Enhanced Independent Component Analysis Approach,” IEEE Transactions on Neural Networks, vol. 18, no. 2,pp.530-541, Mar. 2007

[7] Randa Atta and Mohammad Ghanbari,” Low-Memory Requirement and Efficient Face Recognition System Based on DCT Pyramid,” IEEE Transactions on Consumer Electronics, vol. 56, no. 3, pp.1542-1548,Aug. 2010

[8] Sang-Il Choi, Member and Gu-Min Jeong,” Shadow Compensation Using Fourier Analysis With Application to Face Recognition,” IEEE Signal Processing Letters, vol. 18, no. 1,pp.23-26, Jan.2011

[9] 眼睛圖片,http://big5.sj33.cn/jc/pmjc/ps/200810/17637.html

[10]何政男，2011，高效能視訊文字處理演算法與晶片設計，國立高雄第一科技大學，碩士論文
[11]楊文欽，2010，基於人臉辨識之背光補償調適處理演算法，國立高雄第一科技大學，碩士論文

[12]人臉資料庫，http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html

[13]平板電腦與筆記型電腦圖，http://www.asus.com/tw/

[14] RGB轉YUV公式參考，http://zh.wikipedia.org/wiki/YUV

[15] Dan Witzner Hansen and Arthur E.C. Pece” Iris Tracking with Feature Free Contours,” Proceedings of the IEEE International Workshop Modeling of Faces and Gestures (AMFG’03) , PP. 208-214,2003

[16] Ragini Choudhury Verma, Cordelia Schmid, and Krystian Mikolajczyk,” Face Detection and Tracking in a Video by Propagating Detection Probabilities,” IEEE Transactions on Pattern Analysis And Machine in Telligence, vol. 25, no.10,pp.1215-1228, Oct.2003

[17] Yan Tong, Wenhui Liao and Qiang Ji,” Facial Action Unit Recognition by Exploiting Their Dynamic and Semantic Relationships,” IEEE Transactions on Pattern Analysis And Machine in Telligence, vol. 29, no. 10, pp.1683-1699,Oct.2007

[18] Liya Ding and Aleix M. Martinez,” Features versus Context: An Approach for Precise and Detailed Detection and Delineation of Faces and Facial Features,” IEEE Transactions on Pattern Analysis And Machine in Telligence, vol. 32, no. 11,pp.2022-2038, Nov.2010

[19] Weilong Chen, Meng Joo Er and Shiqian Wu,” Illumination Compensation and Normalization for Robust Face Recognition Using Discrete Cosine Transform in Logarithm Domain”, IEEE Transactions on Systems, Man, And Cybernetics—Part B: Cybernetics, vol. 36, no. 2,pp.458-466, Apr. 2006

[20] Ioana Bacivarov, Mircea Ionita and Peter Corcoran,” Statistical Models of Appearance for Eye Tracking and Eye-Blink Detection and Measurement,” IEEE Transactions on Consumer Electronics, vol. 54, no. 3,pp.1312-1320, Aug.2008

[21] Zhiming Liu and Chengjun Liu,” A Hybrid Color and Frequency Features Method for Face Recognition,” IEEE Transactions on Image Processing, vol. 17, no. 10, pp.1975-1980,Oct. 2008

[22] Harin Sellahewa and Sabah A. Jassim,” Image-Quality-Based Adaptive Face Recognition,” IEEE Transactions on Instrumentation And Measurement, vol. 59, no. 4,pp.805-813, Apr. 2010


[23] William Robson Schwartz, HuiminGuo,Jonghyun Choi and Larry S. Davis,” Face Identification Using Large Feature Sets,” IEEE Transactions on Image Processing, vol. 21,no. 4,pp.2245-2255, Apr. 2012

[24]井上誠喜；八木伸行；林正樹；中須英輔；三谷公二；奧井誠人，《C語言數位影像處理》，全華，民94。

[25]Alasdair McAndrew，《數位影像處理》，初版三刷，高立，台北，民97。

[26] 陳同孝；黃國峰；張真誠，《數位影像處理技術》，旗標，民92

[27] 求是科技，《Visual C++數位影像處理技術大全》，文魁資訊，2008

[28]FreeFileSync軟體，http://freefilesync.sourceforge.net/