臺灣博碩士論文加值系統

平面顯示器是現今顯示器的主流。為了增加液晶顯示器的附加價值，一個可能的方法是增加影像的品質，特別是在色彩表現方向。LED顯示器或以LED背光的顯示器在顏色方面有高飽和度的優點。高飽和度顯示器可以顯示色彩傳真與吸引人的畫面。但是，由於這種顯示器與目前電視或電腦顯示器的標準並不相容，標準視訊在輸入高飽和度顯示器前必須加以修改或轉換。我們考慮一個比標準顯示器色域大的四原色平面顯示器。標準顯示器的色域在本論文中被映射到四原色顯示器的色域。在標準視訊轉換時，保持原色彩的明度與色調，但增加其飽和度。由於轉換時的數值運算是非線性的而且複雜，因此我們使用類神經網路實現這個轉換，使得數值運算可以簡化，並且使得即時運算能夠實現。其他使用類神經網路的優點是改修色域對應演算法時，可以在相同的網路硬體架構下完成，因此有應用上的彈性。因此色差差距堆疊倒傳遞網路被用來訓練網路，其中訓練色塊為512個格子點的色塊。另一組為隨機取樣512個色塊這此被用於測試網路的品質。結果顯示，從標準的訊號轉換到四色原平面顯示器CIELAB座標，對於訓練色塊的平均及最大色差分別為=0.126及0.628。在相同的情形，測試樣品的平均及最大色差分別為=0.748及5.31。

Flat panel displays are popular nowadays. For increasing their add-on values, a possible approach is to increase image quality, especially the color appearance. LED displays or LED-backlight displays have the advantage of high saturation color characteristics. High saturation displays are able to show more faithful and attractive image. However, because they are not compatible with television or display standard, the standard video signals have to be modified or transformed before driving the high saturation displays. We consider a four-primary LED display that its color gamut is larger than standard display. The gamut of the standard display is mapped to the gamut of the four-primary LED display. The lightness and hue are maintained when the standard video signals are transformed. As the numerical operation of the transformation is nonlinear and complicated, we use neural networks to implement the transformation so that the numerical operation can be simplified and real time signal processing is possible. The other advantage of using neural network is the flexibility to alter the gamut mapping algorithm within the same framework of the network hardware. The color-difference stack-enhanced Back-propagation method is used to training the network with 512 grid colors. The other data set of 512 random samples is used to test the performance of the neural network. The results show that, the average and maximum color difference of the training samples are = 0.126 and 0.628, respectively, for the transformation from the standard signals to the CIELAB coordinates of the four-primary display. For the same case, the average and maximum color difference of the test samples are = 0.748 and 5.31, respectively.

目 錄
中文摘要......................... i
Abstract........................ ii
誌 謝 ........................ iii
目 錄 ........................ iv
表目錄........................... v
圖目錄........................... vi
第一章 緒論....................... 1
1-1 引言......................... 1
1-2 研究目的...................... 2
第二章 色彩度量學.................. 3
2-1 色彩度量學的目的............... 3
2-2 色彩度量學的由來與發展歷史...... 3
2-3 CIE三刺激值（tristimulus）的計算... 5
2-4 Lab色彩空間.................... 7
2-5 色域空間轉換................... 15
第三章 類神經網路.................. 21
3-1 網路架構...................... 22
3-2 倒傳遞網路學習演算法........... 25
第四章 實驗裝置與測量.............. 29
4-1 實驗步驟：基本架構的建立....... 29
4-2 倒傳遞神經網路訓練的方法....... 33
第五章 實驗數據與分析.............. 35
5-1 類神經網路收斂性.............. 35
第六章 結論....................... 47
參考文獻......................... 49

[1]羅梅君,“印刷色度學”,印刷科技出版社,1991.
[2]江宗錫,“利用倒傳遞類神經網路作數位相機色彩非線性演繹模式之建立”,元智大學電機與資訊工程研究所碩士論文,2000.
[3]Gunter Wyszecki, W. S. Stiles, “Color Scie Concepts and Methods Quantitative Data and Formulae, 2nd Edition” ,1983.
[4]Mark D. Fairchild, “ Color Appearance Models”, Addison-Wesley Publishing Company, 1998.
[5]Mike Stokes, Mark D. Fairchild, Roy S. Berns,“Precision Requirements for Digital Color Reproduction”, AcM Transactions on Graphics Vol.11,No.4,406-422(1992)
[6]Michael Has, Todd Newman, “Color Management: Current Practics and The Adoption of a New Standard” ,http://www.color.org/.
[7]Mahy, M., L. Van Eyckdenm A. Oosterlinck, “ Evaluation of Uniform Color Spaces developed after the adoption of CIELAB and CIELUV”, Color Research & Application, Vol. 19, No. 2, 105-121(1994).
[8] P. C. Hung,“Colorimetric Calibration in Electronic Image Devices Using Looking-UP-Table Model and Interpolations”,Journal of Electronic Imaging, Vol.2(I),pp.53-61,1993.
[9]Henry R. Kang,“Color Technology For Electronic Image Devices”, SPIE-The International Society for Optical Engineering.
[10]D. I. Specht,“Probabilistic Neural Network for Classification, Mapping, or Associative Memory”,Proceeding of the IEEE International Conference on Neural Networks, Vol,1, pp. 525-532,1988.
[11] F. Rosenblatt, “ The perception: a probabilistic model for information storage and organization in the brain”, Psychological Review, Vol.65, pp. 386-408, 1958.
[12]M. Minsky and S. Papert, Perceptrons, Cambridge, MA, MIT Press, 1969.
[13]D. E. Rumelhart and J. L. McClelland, eds., Parallel Distributed Processing: Exploration in the Microstructure of Cognition,Vol. 1, Cambridge, MA, MIT Press, 1986.
[14]葉怡成, “類神經網路模式應用與實作”,儒林圖書公司,2003.
[15]葉怡成, “應用類神經網路”,儒林圖書公司,1995.