臺灣博碩士論文加值系統

身份認證在軍事及日常生活的安全監控中扮演著相當重要的角色。本研究首先針對人臉進行三維影像拍攝及三維人臉模型建立，並將所獲得的影像進行前處理，辨識人臉之前處理是一項重要步驟，其成果會影響後續辨識的結果。本研究方法主要是先將三維立體影像轉成二維紋理影像，最後以小波相位相關 (Wavelet Phase-Only Correlation, WPOC)為從事二維人臉特徵辨識，用此方式來擷取影像中的特徵。為了改善分類與辨識之效果，我們以WPOC之方法，結合直方圖(Histogram)演算法來進行辨識，我們主要的目的仍在研究如何將三維影像轉成二維影像，且轉換之後的二維影像不會受光照及角度微偏影響辨識率。
本研究先以三維立體影像之特性視深值(Z)軸，並透過剖面方式，轉換成二維紋理影像，最後再以小波相位相關(Wavelet Phase-Only Correlation, WPOC)研究方法為先取出每張圖片之相位差再結合直方圖 (Histogram)進行後續的影像處理，每張影像處理後的結果用二維直方圖方式顯示出來，用判斷直方圖中較明顯峰值出現之分佈情形找出正確的影像，最後本研究將針對WPOC方法、POC方法及主成份分析法(Principal Component Analysis, PCA)進行比較，實驗結果顯示本文辨識率及辨識速度皆高於POC方法及PCA方法，本文利用 100張照片來實驗正確率達99%，辨識速度為0.67秒/張，POC之正確率達97%，辨識速度為1.17秒/張，而PCA之正確率為85%，辨識速度為3.98秒/張。

How to recognizing a person from the face has been gain lots of interests in many applications especially in the security system. It is well known that the 2D face recognition score is low when the face is not in the front views direction. In order to solve this inherent problem of 2D facial image recognition, 3D recognition algorithms have been rigorously developed in these years. Although the 3D image can be recognized in different viewing angles but it has an inherent problem with large amount of data. So that the main issue of 3D image recognition is how to use less data points to achieve the goals.
We try to project the 3D image on the depth direction to form a 2D texture image. This 2D image is similar to the contour lines map as in the geography terminology. We will combine the Wavelet Phase-only Correlation (WPOC) algorithm with Histogram Detection to recognize the facial images. We will inspect the peak distributions in the 2D histogram to find the match image.
The simulation results obtained from the proposed WPOC algorithms will be compared with POC and the Principal Component Analysis (PCA) methods to verify its efficiency. We build our own 3D database from the LT3D face camera. We use totally 100 pictures to test the proposed method. The recognition rate approach 99% with 0.67 second/frame with WPOC. The recognition rate approach 97% with 1.17 second/frame with POC. The recognition rate approach 85% with 3.98 second/frame with PCA.

誌謝 ii
摘要 iii
ABSTRACT iv
目錄 v
表目錄 viii
圖目錄 ix
1. 緒論 1
1.1 研究動機與目的 1
1.2 文獻探討 3
1.3 研究系統流程 4
1.4 論文架構 6
2. 影像辨識基礎理論 7
2.1 結構光掃描編碼方式 7
2.1.1 空間編碼法(Spatial Neighborhood Coding) 7
2.1.2 指向編碼法(Direct Coding) 10
2.1.3 時間多工編碼法(Time-multiplexing Coding) 11
2.2 三維影像拍攝之結構光儀器介紹 13
2.3 色彩編碼與解碼介紹 14
2.4 三維影像之紋理等高線 15
2.4.1 三維影像之輪廓線 16
2.4.2 三維影像之面部曲線 16
2.5 色彩空間 17
2.6 邊緣偵測 17
2.7 形態學 18
2.7.1 膨脹(Dilation) 19
2.7.2 侵蝕(Erosion) 19
2.7.3 斷開(Opening) 19
2.7.4 閉合(Closing) 20
2.8 標記連通元件 20
3. 研究方法與步驟 24
3.1 三維立體模型拍攝與建立 24
3.2 三維立體模型之轉換 25
3.3 有興趣區域(ROI) 26
3.4 離散小波轉換(Discrete Wavelet Transformation, DWT) 27
3.5 傅立葉轉換(Fourier Transform) 30
3.5.1 一維傅立葉轉換對 30
3.5.2 二維傅立葉轉換對 31
3.6 主成份分析法(Principle Components Analysis, PCA) 之介紹 32
3.7 相位相關(Phase-Only Correlation, POC) 34
3.8 小波相位相關(Wavelet Phase-Only Correlation, WPOC) 34
3.9 特徵比對方法：歐幾里得距離（Euclidean Distance） 35
3.10 直方圖之比較方法 35
4. 實驗流程與結果 41
4.1 實驗處理與ROI區域擷取 41
4.2 PCA實驗結果 53
4.3 POC實驗結果 54
4.4 WPOC實驗結果 58
4.5 實驗辨識結果 78
5. 結論 80
參考文獻 81
自傳 84

[1]Gaile, G. G., “Face Recognition Based on Depth and Curvature Features”, IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Proceedings CVPR '92, pp. 808-810, 1992.
[2]Samir, C., “An Intrinsic Framework for Analysis of Facial Surfaces”， International Journal of Computer Vision, pp.80-95, 2009.
[3]Chafik, S., Anuj, S., and Mohamed, D., “Automatic 3D Face Recognition Using Shapes of Facial Curves”, IEEE International Conference on Acoustics, Speech and Signal Processing, pp. 1-10, 2006.
[4]Pinto, S. C. D., “3D Facial Expression Analysis By Using 2D and 3D Wavelet Transforms”, IEEE International Conference on Image Processing (ICIP), pp. 1281-1284, 2011.
[5]Huibin, L., Pierre, H. Di., Jean-Marie, L. M., and Liming, C., “Expression Robust 3D Face Recognition Via Mesh-based Histograms of Multiple Order Surface Differential Quantities”, IEEE International Conference on Image Processing (ICIP), pp. 3053-3056, 2011.
[6]Vretos, N., “3D Facial Expression Recognition Using Zernike Moments on Depth Images”, IEEE International Conference on Image Processing (ICIP), pp. 773-776, 2011.
[7]Fang, T., “4D Facial Expression Recognition”, IEEE International Conference on Computer Vision Workshops (ICCV Workshops), pp. 1594-1601, 2011.
[8]Choi, J., “Real-time 3-D Face Tracking and Modeling From a Webcam”, IEEE Workshop on Applications of Computer Vision (WACV), pp. 33-40, 2012.
[9]Agarwal, M., “Face Recognition using Principle Component Analysis”, International Conference on Signal Acquisition and Processing, pp. 310-314, 2010.
[10]Caspi, D., “Range Imaging With Adaptive Color Structured Light”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 20, pp. 470-480, 1998.
[11]Zhang, L., “Rapid Shape Acquisition Using Color Structured Light and Multi-pass Dynamic Programming”, In Proceedings of the 1st International Symposium on 3D Data Processing, Visualization, and Transmission (3DPVT), pp. 34-36, 2002.
[12]Fechteler, P., “Fast and High Resolution 3D Face Scanning”, IEEE International Conference on Image Processing, pp. 81-84, 2007.
[13]Tehrani, M. A., “A New Approach to 3D Modeling Using Structured Light Pattern”, International Conference on Information and Communication Technologies, pp. 1-5, 2008.
[14]Jason, G., “Structured-light 3D surface imaging: a tutorial”, Advances in Optics and Photonics, Vol. 3, Issue 2, pp. 128-160, 2011.
[15]Besl, P. J., “A Method for Registration of 3-D Shapes”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 14, pp. 239-256, 1992.
[16]Qiang, Z., “Discriminant Improved Local Tangent Space Alignment with Adaptively Weighted Complex Wavelet for Face RecognitionA”,IEEE International Conference on Image Processing, pp. 3708-3713, 2012.
[17]Wang, L., “Face Detection Based on QFT Phase-Only Correlation Template Match”, Proceeding ICISE '09 Proceedings First IEEE International Conference on Information Science and Engineering, pp. 1231-1234, 2009.
[18]Basit, F., “Face Recognition using Processed Histogram and Phase-Only Correlation (POC)"”, IEEE International Conference on Image Processing, pp. 238-242, 2007.
[19]Maruyama, M., “Range sensing by projecting multiple slits with random cuts”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 20, pp. 647-651, 1993.
[20]Durdle, N. G., “An Improved Structured Light Technique for Surface Reconstruction of the Human Trunk”, IEEE Canadian Conference on Electrical and Computer Engineering, pp. 874-877, 1998.
[21]Vuylsteke, P., “Range Image Acquisition with a Single Binary-Encoded Light Pattern”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 12, pp. 148-164, 1990.
[22]Cheng, F., “3D Object Scanning System by Coded Structured Light”, Third International Symposium on Electronic Commerce and Security (ISECS), pp. 213-217, 2010.
[23]http://www.cs.cmu.edu/afs/cs/academic/class/15385-s06/lectures/ppts/lec-17.ppt (2012.5.1)
[24]http://www.ltech.com.tw/ltech/about/about_company.php (2012.5.1)
[25]Michael, W. S., William, B. C., and John, C. B., “An Experimental Comparison of RGB, YIQ, LAB, HSV, and Opponent Color Models” , ACM Transactions on Graphics, Vol. 6, no. 2, pp. 123-158, 1987.
[26]繆紹綱，數位影像處理，普林斯頓國際有限公司，第364-392頁，2007。
[27] Basart, J. P. and Gonzalez, R. C., “Binary Morphology”, in Advances in Image Analysis, Y. Mahdavieh and R. C. Gonzalez (eds.), SPIE Press, Bellingham, Wash, pp. 277-305, 1992.
[28]Basart, J. P., Chacklackal, M. S., and Gonzalez, R. C., “Introduction to Gray-Scale Morphology”, in Advances in Image Analysis, Y. Mahdavieh and R. C. Gonzalez (eds.), SPIE Press, Bellingham, Wash, pp. 306-354, 1992.
[29]邱柏智，“基於主成份分析法與灰關聯分析法之人臉辨識”，第一屆電資科技應用與發展學術研討會，第1卷，第1-15頁，2007。