近年來，膚色偵測是電腦視覺研究裡一門相當重要的技術之一，且膚色偵測已廣泛應用在許多人臉偵測或自動監控系統中。但是光源的影響一直是膚色偵測技術中主要挑戰，在不同光源照射下，皮膚所反應的顏色也會有所改變，這會導致系統誤判率提高。因此，本論文提出一個以高斯混合模型為基礎之色彩恆常性演算法，我們利用高斯混合模型建立不同光源下色域的機率分佈，並找出影像中之光源並校正該影像以提高膚色偵測正確率。於本實驗中，實驗使用「A Data Set for Color Research」資料庫、CMU PIE人臉資料庫及自行建立的人臉影像，並透過合成光源法產生論文實驗影像。在實驗結果裡本論文所提出方法能夠有效的校正影像，且抵抗不同背景的影響。

Computer vision has become one of the most important research topics for intelligent system. Skin color detection is one critical approach to detect objects due to its robustness to rotation, scale change, shape variation, and occlusion. The challenge of skin detection is that the skin appearance in images is easily affected by various illuminants. In this paper, a color constancy method is proposed for skin color detection. The method represents the probability distributions of color gamuts of different illuminants find the illuminant of an image by computing correlation of color gamuts, and correct the image by compensating the color gain of the illuminant. In the experiments, we use the “A Data Set for Color Research database”, “CMU PIE face database” and the face images captured by ourselves. Synthesized-illuminated images are used to train the GMM color constancy method. In the experiment results, our method can correct the image effectively compared with gray world, white patch, and color by correlation methods, and robust to environmental changes.

摘要.....................................................................................................................................i
英文摘要.............................................................................................................................ii
誌謝...................................................................................................................................iii
目錄...................................................................................................................................iv
表目錄................................................................................................................................v
圖目錄................................................................................................................................vi
第1 章 簡介.......................................................................................................................1
1.1 研究背景..................................................................................................................1
1.2 研究動機..................................................................................................................4
1.3 論文架構..................................................................................................................5
第2 章 傳統色彩恆常性演算法.......................................................................................7
2.1 灰界假設演算法(Gray World Assumption，GWA) ..............................................7
2.2 白色區塊法(White Patch)........................................................................................8
2.3 Retinex 演算法.........................................................................................................9
2.4 色域對應演算法(Gamut Mapping Algorithm，GMA) ........................................11
第3 章 GMM Color Constancy .......................................................................................12
3.1 Color by correlation 演算法....................................................................................12
3.2 高斯混合模型(Gaussian Mixture Model，GMM) ...............................................15
3.2.1 期望最大演算法(Expectation-Maximization Algorithm，EM) ....................16
第4 章 具抗光源變化的膚色偵測.................................................................................18
第5 章 實驗結果.............................................................................................................22
5.1 色彩恆常性方法比較............................................................................................25
5.1.1 GMM color constancy 索引色域的訓練.........................................................25
5.1.2 索引色域色度空間的比較.............................................................................26
5.2 不同光源下人臉影像中的膚色偵測....................................................................27
5.2.1 GMM color constancy 索引色域的訓練.........................................................28
5.2.2 GMM 膚色分類器的訓練...............................................................................32
5.2.3 膚色偵測結果的評估.....................................................................................34
5.2.4 不同色彩恆常性方法的膚色偵測比較.........................................................35
第6 章 結論.....................................................................................................................41
參考文獻...........................................................................................................................42

[1]P. Kakumanu, S. Makrogiannis, and N. Bourbakis, “A survey of skin-color modeling and detection methods,” Pattern Recognition, Vol. 40, Issue 3, pp. 1106-1122, 2007.
[2]M. Ebner, Color constancy, Wiley, 2007.
[3]H.-C. Lee, Introduction to color imaging science, Cambridge University Press, 2005.
[4]S. Westland and C. Ripamonti, Computational colour science using Matlab, John Wiley & Sons, 2004.
[5]R. Gershon, A. D. Jepson, and J. K. Tsotsos, “From [R,G,B] to surface reflectance: computing color constant descriptors in images,” Proceedings, Internation Joint Conference on Artificial Intelligence, Milan, Italy, pp. 755-758, 1987.
[6]G. Buchsbaum, “A spatial processor model for object color perception,” Journal of the Franklin Institute, Vol. 310, No. 126, pp. 1-26, 1980.
[7]K. Barnard, V. Cardei, and B. Funt, “A comparison of computational color constancy algorithms—Part I: methodology and experiments with synthesized data,” IEEE Transactions on Image Processing, Vol. 11, Issue 9, pp. 972-983, 2002.
[8]K. Barnard, L. Martin, A. Coath, and B. Funt, “A comparison of computational color constancy algorithms—Part II: experiments with image data,” IEEE Transactions on Image Processing, Vol. 11, Issue 9, pp. 985-995, 2002.
[9]C. Rosenberg, T. Minka, and A. Ladsariya, “Bayesian color constancy with non-Gaussian models,” Proceedings, Neural Information Processing Systems (NIPS), British Columbia, 2003.
[10]A. Rizzi, C. Gatta, and D. Marini, “A new algorithm for unsupervised global and local color correction,” Pattern Recognition Letters, Vol. 24, Issue 11, pp. 1663-1677, 2003.
[11]E. H. Land, “The Retinex,” American Scientist, Vol. 52, No. 2, pp. 247-264, 1964.
[12]E. H. Land, “The Retinex theory of colour vision,” Proceedings, Royal Institution of Great Britain, London, Britain, pp. 23-58, 1974.
[13]E. H. Land, “An alternative technique for the computation of the designator in the retinex theory of color vision,” Proceedings, National Academy of Science of the United States of America, pp. 3078-3080, 1986.
[14]B. Funt, F. Ciurea, and J. McCann, “Retinex in Matlab,” Proceedings, IS&T/SID Color Imaging Conference: Color Science, Systems and Applications, pp. 112-121, 2000.
[15]B. Funt, F. Ciurea, and J. McCann, “Retinex in Matlab,” Journal of Electronic Imaging, Vol. 13, No. 2, pp. 48-57, 2004.
[16]Z. Rahman, D. J. Jobson, and G. A. Woodell, “Multi-scale retinex for color image enhancement,” Proceedings, International Conference on Image Processing, Lausanne, Switzerland, pp. 1003-1006, 1996.
[17]D. A. Forsyth, “A novel approach to colour constancy,” Proceedings, Second International Conference on Computer Vision, Tampa, USA, pp. 9-18, 1988.
[18]D. A. Forsyth, “A novel algorithm for color constancy,” International Journal of Computer Vision, Vol. 5, Issue 1, pp. 5-36, 1990.
[19]G. D. Finlayson, S. D. Hordley, and P. M. Hubel, “Color by correlation: a simple, unifying framework for color constancy,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 23, Issue 13, pp. 1209-1221, 2001.
[20]S. Akaho, “Mixture model for image understanding and the EM algorithm,” TR-95-13, Electrotechnical Laboratory, Umezono, Ibaraki, Janpan, 1995.
[21]J. A. Bilmes, “A gentle tutorial of the EM algorithm and its application to parameter estimation for Gaussian mixture and hidden Markov models,” TR-97-021, International Computer Science Institute, Berkeley, CA, USA, 1998.
[22]T. K. Moon, “The expectation-maximization algorithm,” IEEE Signal Processing magazine. Vol. 13, Issue 6, pp. 47-60, 1996.
[23]F. Dellaert, “The expectation maximization algorithm,” GIT-GVU-02-20, Georgia Institute of Technology College of Computing, Georgia, USA, 2002.
[24]G. Verbeke and G. Molenberghs, Linear mixed models for longitudinal data, Springer, 2000.
[25]D. J. C. MacKay, Information theory, inference, and learning algorithms, Cambridge University Press, 2003.
[26]G. Singh, A. Panda, S. Bhattuchulyya, and T. Stikunthun, “Vector quantization techiques for GMM based speaker verification,” Proceedings, IEEE International Conference on Acoustics, Speech, and Signal Processing. (ICASSP '03), Hong Kong, China, pp. 6-10, 2003.
[27]S. Nasser, R. Alkhaldi, and G. Vert, “A modified Fuzzy K-means clustering using expectation maximization,” Proceedings, IEEE International Conference on Fuzzy Systems, BC, Canada, 2006.
[28]S. L. Phung, A. Bouzerdoum, and D. Chai, “Skin segmentation using color pixel classification: analysis and comparison,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 27, Issue 1, pp. 148-154, 2005.
[29]K. Barnard, L. Martin, B. Funt, and A. Coath, “A data set for color research,” Color Research and Application, Vol. 27, pp. 148-152, 2002.
[30]T. Sim, S. Baker, and M. Bsat, “The CMU pose, illumination, and expression database,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 25, Issue 12, pp. 1615-1618, 2003.
[31]H. L. V. Trees, Detection, estimation, and modulation theory, part I: detection, estimation, and linear modulation theory, Wiley, 2001.
[32]M. J. Jones and J. M. Rehg, “Statistical color models with application to skin detection,” Proceedings, IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Fort Collins, CO, USA, pp. 274-280, 1999.