在拍照的過程中，閃光燈被當作場景的第二光源來使用。然而，當閃光燈的色彩和環境光的色彩有所不同時，將產生顏色不一致的問題。在我們的研究中，將試著找出閃光燈影像上每一點的光源色彩，進而校正這些色彩至相同的顏色。我們的方法是基於影像本質物件的分解 (Intrinsic image decomposition)。它用來描述觀察到的影像和光源色彩之間的關係。在本篇論文中，我們提出一套二段式的分解流程，包括全域性光源色彩估計以及局部性光源修正。首先，我們分別對閃光燈和場景光的色彩估計，找出全域性的光源色彩分布；接著，藉由找出場景物體的主要色彩來修正前一階段求出的光源色彩分布。這套分解流程在結合前景物體的偵測後，可以被延伸至自動化的型式。從實驗結果顯示，我們提出的自動化流程在使用我們所建立的閃光燈影像資料庫中，能得到與現有半自動的方法近似的結果。另外，我們也討論了這個方法的限制以及日後的改善空間。最後，我們總結我們的方法，它能自動化解決閃光燈影像中的色彩不一致問題，並提供良好的校正結果。

Flashlight is used as the second illumination of a scene in photography. There exists a color inconsistent problem in images once the light color of flashlight and the light color of scene are different. The purpose of our work is to find the light color of each pixel in flashlight images, and then adjust the light color of each pixel to be the same. Our work is based on intrinsic image decomposition. The intrinsic image model the relationship between the observed image and the light color. In the thesis, we propose a two-stage decomposition method including global light map estimation and local light map refinement. First, we calculate the light map by finding the two light colors of both flashlight and the scene. Then we refine the light map by finding the dominative object’s colors. This decomposition method can be further extended to automatic form by combining with foreground object detection method. The experimental results show that the performance of our proposed automatic method is satisfactory in comparison with existing semi-automatic correction methods with our flashlight database. In addition, we also discuss the limitation and future work of our approach. In summary, the proposed method automatically solves the color inconsistent problem in flashlight images and also provides promising results.

[1] Y. Weiss, “Deriving Intrinsic Images from Image Sequences,” Proc. ICCV, 2001.
[2] E. Eisemann and F. Durand, “Flash Photography Enhancement via Intrinsic Relighting,” Proc. SIGGRAPH, 2004.
[3] E. Hsu, T. Mertens, S. Paris, S. Avidan and F. Durand , ”Light Mixture Estimation For Spatially Varying White Balance,” Proc. SIGGRAPH, 2008.
[4] A. Bousseau, S. Paris, and F. Durand, “User-Assisted Intrinsic Images,” Proc. SIGGRAPH Asia, 2009.
[5] H. G. Barr and J. M. Tenenbaum, “Recovering Intrinsic Scene Characteristics from Images,” Computer Vision Systems, pp. 3-26, 1978.
[6] G. Buchsbaum, "A spatial processor model for object color perception", Journal of the Franklin Inst., vol. 310, pp. 1 - 26, 1980.
[7] J. Weijer and T. Gevers, “Color Constancy based on the Grey-Edge Hypothesis,” Proc. ICIP, 2005
[8] R. Lu, A. Gijsenij. T. Gevers, K. Sande, J. Geusebroek, and D. Xu, “Color Constancy using Stage Classification,” Proc. ICIP, 2009.
[9] D. A. Forsyth, “A Novel Algorithm for Color Constancy,” International Journal of Computer Vision, vol. 5(1), pp. 5-36, 1990.
[10] G. D. Finlayson and S.D. Hordley, “Gamut Constrained Illuminant Estimation,” International Journal of Computer Vision, vol. 67(1), pp. 93-109, 2006.
[11] A. Gijsenij, T. Gevers, and J. van de Weijer, “Generalized Gamut Mapping using Image Derivative Structures for Color Constancy,” International Journal of Computer Vision, vol. 86(2–3), pp. 127-139, 2010.
[12] E. H. Land and J. J. McCann, “Lightness and Retinex Theory,” Journal of the Optical Society of America, 61(1):1-11, 1978.
[13] R. Kimmel, M. Elad, D. Shaked, R. Keshet, and I. Sobel, “A Variational Framework for Retinex,” International Journal of Computer Vision, vol. 52(1), pp. 7-23, 2003.
[14] M. F. Tappen, W. T. Freeman, and E. H. Adelson, ”Recovering Intrinsic Images from a Single Image,” IEEE Trans. PAMI, Sep 2005.
[15] B. Funt and F. Ciurea, “Retinex in MATLAB,” Proceedings of the IS&T/SID Eighth Color Imaging Conference: Color Science, Systems and Applications, pp. 112-121, 2000.
[16] M. Ebner, “Color Constancy Using Local Color Shifts,“ Proc. ECCV, 2004
[17] L. Shen, P. Tan, and S. Lin, “Intrinsic Image Decomposition with Non-Local Texture Cues,” Proc. CVPR, 2008.
[18] A. Levin, D. Lischinski, and Y. Weiss, “A Closed Form Solution to Natural Image Matting,” IEEE Trans. PAMI, Feb 2008.
[19] F. Ciurea and B. Funt, “A Large Image Database for Color Constancy Research,” proceedings of the eleventh color imaging conference, pp. 160-164, 2003.
[20] http://en.wikipedia.org/wiki/Color_temperature
[21] http://www.bobatkins.com/photography/digital/canon_eos_30D_specifications.html
[22] http://photonotes.org/cgi-bin/flash-specs.pl?flashType=canon430ex
[23] http://en.wikipedia.org/wiki/Black_body
[24] http://en.wikipedia.org/wiki/Planck%27s_law
[25] SaliencyToolbox. http://www.saliencytoolbox.net/
[26] OpenCV Viola-Jones Face Detection in Matlab.
http://www.mathworks.com/matlabcentral/fileexchange/19912
[27] http://en.wikipedia.org/wiki/Jaccard_index
[28] OpenCV 2.1.
http://www.willowgarage.com/blog/2010/04/12/opencv-21-released
[29] K. Barnard, L. Martin, B. Funt, and A. Coath, “A Data Set for Colour Research,” Color Research and Application, Vol. 27, No. 3, pp. 147-151, 2002.
[30] R. Grosse, M. K. Johnson, E. H. Adelson, W. T. Freeman, “Ground truth dataset and baseline evaluations for intrinsic image algorithms,” Proc. ICCV, 2009.