臺灣博碩士論文加值系統

現今市面上各家廠商生產的顯示器並無統一的標準測試規範，顯示器的色域範圍(Gamut)都不一致，所以對於高動態範圍影像(High Dynamic Range Images)在進行跨媒體進行轉換時，就必須針對從可能『屬廣色域範圍』的原稿端到『屬較窄色域範圍』的複製端的成像設備，做最佳化的跨媒體色域對映運算，使影像在具不同大小色域範圍的成像媒體上，均能夠有較佳的色彩呈現效果。在本研究中之所以採用高動態範圍的影像來進行色域對映，乃在於『如何在窄色域範圍的顯示設備上，將高動態影像的細部層次保留下來，以維持其較優良的影像品質，視為相當大的研究挑戰』。
由於在跨媒體色域空間轉換後，輸出端在顯示影像時，得考慮到觀測者的觀測條件會有所不同，所以本研究亦同時考慮到觀測條件的影響，而應用「色外貌模式」進行觀測條件參數的設定，讓影像輸出能有符合觀測者的所有觀測環境因素。
本研究的重點，在於『高動態彩色影像的跨媒體〝色域對映〞』。實驗的步驟分為三個子流程。首先在第一子流程，先建立原稿端和複製端兩者的「設備色彩特性演繹模組」。至於第二子流程，則在於進行「多張影像融合的HDR成像」(High-Dynamic-Range Imaging)與「擬高動態成像」(HDR-like imaging)，以得到高動態的彩色影像；最後的第三子流程，則進行HDR影像跨媒體的優化轉換，首先，先進行色域對映轉換的運算，再以輸出端Hi-Fi印表機的反推型設備色彩特性演繹模式，進行頻譜轉換複製的運算。最終，可以得到近似模擬高動態影像─「能有良好的階調層次與明暗變化」，且具有Hi-Fi表現的軟式打樣彩色影像成品(Soft-Proofing)，達成影像跨媒體複製的最佳化效果。

There is no unified standard display test specification on the market between manufacturers. Every output device’s gamut is different from others. Cross-media transform of HDR images (High –Dynamic-Range Images) needs to be visually reasonable and smooth via gamut mapping algorithms especially from an input device with a bigger color gamut to an output device having smaller gamut. Therefore, the aim of this research is to study the application of gamut mapping on HDR images, because both detail and gradation of HDR image can be satisfactorily rendered on an output/display device having smaller gamut is quite a considerable challenge on cross-media color reproduction systems/market.
Practically, viewing conditions can be disparately different from one another. This research, therefore, also took view conditions into account, and used CIECAM02 color appearance model to adjust output images according to the view condition of interest. As results, the output rendered images could be in compliance with all environmental aspects of the observer.
In this research, the key focus was on cross-media gamut mapping applied on HDR. There were three main steps carried out in this study. First step was the set-up or construction of GBPs (Gamut Boundary Points) for both the input and the output devices considered. The next step was the process of HDR imaging, include a multiple image fusion and a HDR-like imaging. Finally the last step was to optimize the HDR type of cross-media color transform. Two type of output imaging devices were characterized and applied for these testing. One was a soft-proofing of CMYKOG High-Fidelity (Hi-Fi) printer, characterized using a multispectral approach. Another was a display, formatted with the standard of Adobe RGB color space. The experimental results showed that, by these three optimization steps of processing, those soft-copy image, which were rendered from either process of both the multiple-images confusion and HDR-like imaging, could gave visually satisfactory color appearance with pleasing details and gradations.

目錄
摘要 1
Abstract 2
目錄 5
圖目錄 7
表目錄 9
第一章 緒論 10
1.1 研究背景 10
1.1.1 色域對映 10
1.1.2 色外貌模式 10
1.1.3 高動態影像 11
1.2 研究動機與目的 12
1.3 研究範圍 14
第二章 文獻探討 16
2.1 高動態影像融合 16
2.2 高斯金字塔 17
2.3 擬高動態範圍成像技術 19
2.4 色域對映 21
2.4.1 裁切法(Clipping)： 21
2.4.2 線性壓縮法(Linear Compression)： 22
2.4.3 非線性壓縮法(Non-Linear Compression)： 23
2.4.4 單一集中壓縮點(Single Focal Point) 24
2.4.5 複數集中壓縮點(Multiple Focal Point) 25
2.5 肩部或膝部的明度調整 25
2.5.1 膝部明度壓縮調整 26
2.5.2 膝部明度增強調整 27
2.5.3 肩部明度壓縮調整 29
2.5.4 肩部明度增強調整 30
2.6 CIECAM02色外貌模式 32
2.7 Lagrange內插法 35
第三章 研究方法 36
3.1 高動態成像 37
3.2 影像處理流程 39
3.2.1 色域對映 39
第四章 實驗流程、結果與分析 45
4.1 實驗流程 45
4.1.1 高動態成像 45
4.1.2 色域建立 46
4.1.3 色域對映 48
4.2 實驗結果與分析 48
第五章 結論與建議 64
5.1 結論 64
5.2 建議 65
參考文獻 66

Berrut, J. P., & Trefethen, L. N. (2004). Barycentric Lagrange Interpolation. Society for Industrial and Applied Mathematics, pp. 201-517.
2.Frei, W. (1977). Image Enhancement by Histogram Hyperbolization. Computer Graphics and Image Processing 6 pp. 286–294.
3.Hummel, R. A. (1977). Image Enhancement by Histogram Transformation.Computer Graphics and Image Processing 6 pp. 184-195.
4.Lagrange polynomial. (2013). Retrieved from wikipedia: http://en.wikipedia.org/wiki/Lagrange_interpolation
5.Lo, M. C., Chen, Y. L., Chang, C. W., & Chen, Y. L. (2004). The Application of IT8.7/3 Test Target in Color Transformations for Printing Devices via Look-Up Tables. TAGA.
6.Lo, M. C., Lin, J. L., & Li, W. B. (2013). New HDR-type Imaging Methods Based on LCH(CIELAB) and JCH(CIECAM02) Color Spaces.
7.Moroney, N., Fairchild, M. D., Hunt, R. W., Li, C., Luo, M. R., & Newman, T. (2002). The CIECAM02 Color Appearance Model.
8.Pizer, S., Amburn, E. P., Austin, J., & et al., (1987). Adaptive Histogram Equalization and Its Variations. Computer Vision, Graphics, and Image Processing pp. 355-368.
9.Zuiderveld, K. (1994). Contrast Limited Adaptive Histogram Equalization. Graphics Gems IV Academic Press. pp. 474-485.
10.馬欣愷 (2011)。 多張影像融合與階調對映演算法之研究。 世新大學資訊管理研究所碩士論文.
11.闕家彬 (2003)。 色彩管理系統進階色域對映技術。 世新大學平面傳播科技研究所碩士論文.