反光及陰影都是光線變化所造成的，然而這會影響到許多電腦視覺及影像處理的演算法，例如影像分割、物體追蹤、識別、偵測等。因此，發展一套可以將陰影/反光去除的系統是有其必要的。這本篇論文裡，我們將使用多重解析度為基準來分離出單張影像中的陰影/反光。首先我們利用圖形模型來分離出陰影，與先前提出的方法不同的是，訓練資料為不同地點取得的影像序列，並將其分割成小塊的影像，然後用Markov網絡找出之間的關係。將測試資料套用我們訓練出來的關係，即可獲得我們所要的結果。在論文的第二部份，我們先使用一個完全只利用顏色特性的方法去偵測出反光的區域，接著利用多重解析度的特性將反光的部分分離出來。首先把影像降維度，直到影像中沒有反光成分的存在。在做升維度的時候，非反光的區域將被保留，而反光區域由較低解析度的圖更新。我們利用求最小平方誤差解這個方法來修正升維度的結果，最後取得一張沒有反光的影像。

Both shading and specular are caused by lighting, and produce troublesome effects for algorithms of computer vision and image processing, such as segmentation, object tracking, recognition, detection, etc. Therefore, develop a system to separate them out is required. In this paper, we use multi-resolution based methods to solve shading/specular problem using a single image. Firstly, we discuss a method to separate shading and reflectance with graphical mode. Unlike previous methods of estimating shading images, our training data are generated from image sequences shot from different scenes and lightings. Training sets are generated by pairing original image patches with their illumination image patches at different scales, and then their relationships are modeled by Markov network. During testing, belief propagation is utilized to search for illumination patches that best described the input image patches by both considering shading consistency and minimizing the total amount of edge derivates. In the second part, we detect specular by a method solely on colors. Then we do down-sampling until there is no any specular components exists. When up-sampling, we update the specular pixels by lower resolution and keep diffuse pixels. Finally, least square solution is used for correcting our results.

[1] H.G. Barrow, J.M. Tenenbaum, “Recovering Intrinsic Scene Characteristics from Images,” Computer Vision System, pp. 3-26, 1978.
[2] M. Baba, A.B. Smith, and E.F. Roberts, “Shadow Removal from a Real Picture,” Proc. SIGGRAPH Conf. Sketches and Applications: In Conjunction with the 30th Annual Conf. Computer Graphics and Interactive Techniques, pp. 27-31, 2003.
[3] Y.C. Chung, J.M. Wang, R.R. Bailey, S.W. Chen, S.L. Chang, and S. Cherng, “Physical-Based Extraction of Intrinsic Images from a Single Image,” Proc. 17th IEEE Int. Conf. Pattern Recognition, Vol. 4. pp. 693-696, 2004.
[4] R. Fattal, D. Lischinski, M. Werman, “Gradient Domain High Dynamic Range Compression,” Proc. SIGGRAPH, pp. 249-356, 2002.
[5] G.D. Finlayson, M.S. Drew, and C. Lu, “Intrinsic Images by Entropy Minimization,” Europe Conf. Computer Vision, Vol. 3023. pp. 582-595, 2004.
[6] W.T. Freeman, E.C. Pasztor, and O.T. Carmichael, “Learning Low Level Vision,” Int. J. Computer Vision, Vol. 40. No. 1. pp. 25-47, 2000.
[7] Y. Matsushita, K. Nishino, K. Ikeuchi, and M. Sakauchi, “Illumination Normalization with Time-Dependent Intrinsic Images for Video Surveillance,” IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 26. No. 10. pp. 1336-1347, 2004.
[8] J. Pearl, Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference, Morgan Kaufmann, 1988.
[9] M. Tappen, W.T. Freeman, and E.H. Adelson, “Recovering Intrinsic Images from a Single Image,” IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 27. No. 9. pp. 1459-1472, 2005.
[10] M. Tappen, E.H. Adelson, and W.T. Freeman, “Estimating Intrinsic Component Images using Non-Linear Regression,” IEEE Int. Conf. Computer Vision and Pattern Recognition, Vol. 2. pp. 1992-1999, 2006.
[11] Y. Weiss, “Deriving Intrinsic Images from Image Sequences,” Proc. 9th IEEE Int. Conf. Computer Vision, pp. 68-75, 2001.
[12] Y. Chu and Z. Chen, "Shadow detection and removal for still images using multi-resolution gradient information," Computer Vision, Graphics and Image Processing Conference, 2006.
[13] Q. He and C. Chu, "Recovering Intrinsic Images from Weighted Edge Maps," IEEE International Conference on Intelligent Information Hiding and Multimedia Signal, 2006
[14] R.T. Tan, K. Nishino, and K. Ikeuchi, "Illumination Chromaticity Estimation using Inverse-Intensity Chromaticity Space," proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 673-680, 2003.
[15] R.T. Tan and K. Ikeuchi, "Separating Reflection Components of Textured Surface using a Single Image," International Conference on Computer Vision, Vol. 2, pp. 982-987, 2003.
[16] R.T. Tan and K. Ikeuchi, "Intrinsic Properties of an Image with Highlights," proceeding of Meeting on Image Recognition and Understanding, July 23-25, 2004, Hakodate, Japan.
[17] R.T. Tan, K. Nishino, and K. Ikeuchi, "Separating Reflection Components Based on Chromaticity and Noise Analysis," IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 26, No. 10, pp.1373-1379, Oct. 2004.
[18] R.T. Tan and K. Ikeuchi, "Separating Reflection Components Of Textured Surface Using a Single Image," IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 27, No. 2, pp.173-193, Feb. 2004.
[19] R. Feris, R. Raskar, K. Tan, and M. Turk, "Specular Highlights Detection and Reduction with Multi-Flash Photography," Journal of the Brazilian Computer Society, 2006.
[20] A. DelPozo and S. Savarese, "Detecting Specular Surfaces on Natural Images," proceeding of IEEE Conference on Computer Vision and Pattern Recognition, 2007.
[21] L.B. Wolff and T. Boult, “Constraining Object Features Using Polarization Reflectance Model,” IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 13, No. 7, pp. 635-657, July 1991.
[22] S.K. Nayar, X.S. Fang, and T. Boult, “Separation of Reflection Components Using Color and Polarization,” Int’l J. Computer Vision, Vol. 21, No. 3, 1996.
[23] Y. Sato and K. Ikeuchi. Temporal-color space analysis of reflection. Journal of Optics Society of America A., Vol. 11, No. 11, pp. 2990–3002, 1994.
[24] S. Lin and H.Y. Shum. Separation of diffuse and specular reflection in color images. In proceeding of IEEE Conference on Computer Vision and Pattern Recognition, Vol. 1, pp. 341–346, 2001.
[25] S. Shafer. Using color to separate reflection components. Color Research and Applications, Vol. 10, pp.210–218, 1985.
[26] H.C. Lee, E.J. Breneman, and C.P. Schulte. Modeling light reflection for computer color vision. IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol. 12, No. 4, pp.402–409, 1990.