立體匹配演算法中，可以被大略的分成兩大類的方法，一種為全域演算法，需要做能量最佳化來決定每個點的視差值；一種則是區域演算法，每個點選擇各自的最佳解即可。雖然方法被分成兩大類，但是要使最後的深度圖精準的話，最重要的前提依然是成本的建構方式，我們甚至可以說如果成本的計算上夠準確的話，不論用哪一大類的方法都可以得到相當不錯的結果。
所以在本篇論文中，我們提出一個新的成本建構方式，結合特徵描述式和新的視窗演算法來建構成本。這個成本建構方法可用來解決立體匹配上容易發生錯誤的地方，如:遮蔽處、紋理重覆處、平滑處。
我們所提出的演算法分成四個部分，第一部分:成本的建構方式，採納顏色和特徵描述式和新的視窗演算法；第二部分:成本的聚合方式，使用十字演算法來生成每個點的支持區域，並聚合每個區域的成本，以用來得到一個穩定值；第三部分:使用掃描線最佳化，來優化每個點的成本，以得到一個更精準的深度圖；第四部份:使用各種檢測方式，得到圖中不可信賴的點，再使用各種不同的優化演算法，對這些點優化，得到最終的深度圖。

In the existing stereo matching methods, different definitions of energy functions are proposed and minimized to obtain depth values of each pixel. Those energy functions can be approximately divided into two terms: the data term and the smooth term. In most of methods, the data term is considered and usually dominate the accuracies of the depth values. However, occlusions, repeated textures and smooth regions in images would comprise the performances of the data term.
Therefore, in this paper, we present a novel definition of a data term which considers pixel colors, image descriptions of each pixel and multiple occluding situations. The image description is used to overcome the repeated textures and the smooth regions by matching wider regions in left and right images without increasing the computational complexities. An occluding detection window is utilized to recognize different occluding situations and give appropriate cost estimations.
Finally, we use cost aggregation methods to verify costs of each pixel. Refine methods are also applied to smooth discontinuities of depth maps. The experiment shows compelling performances using numbers of benchmarks and comparing with ground truths.

[1] W.-Y. Chen, Y.-L. Chang, S.-F. Lin and L.-F. Di, "Efficient Depth Image Based Rendering with Edge Dependent Depth Filter and Interpolation," IEEE International Conference on Multimedia and Expo., pp. 1314-1317, 2005.
[2] Y. Deng, Q. Yang, X. Lin and X. Tang, "Stereo Correspondence with Occlusion Handling in a Symmetric Patch-Based Graph-Cuts Model," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 29, no. 6, pp. 1068 - 1079, 2007.
[3] G. Chen and L. Hong, "Segment-based stereo matching using graph cuts," IEEE Conference on Computer Vision and Pattern Recognition, vol. 1, pp. I-74-I-81, 2004.
[4] Y. Boykov, O. Veksler and R. Zabih, "Fast approximate energy minimization via graph cuts," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 23, no. 11, pp. 1222-1239, 2001.
[5] Q. Yang, L. Wang, R. Yang and H. Stewen, "Stereo Matching with Color-Weighted Correlation, Hierarchical Belief Propagation, and Occlusion Handling," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. vol.31, no. 3, pp. 492-504, 2009.
[6] J. Sun, N.-N. Zheng and H.-Y. Shum, "Stereo Matching Using Belief Propagation," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 7, pp. 787-800, JULY 2003.
[7] O. Veksler, "Stereo correspondence by dynamic programming on a tree," IEEE Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 384-390, 2005.
[8] A. Hosni, B. M., G. M. and C. Rhemann, "Local stereo matching using geodesic support weights," IEEE International Conference on Image Processing, pp. 2093-2096, 2009.
[9] X. Sun, X. Mei, S. Jiao, M. Zhou and H. Wang, "Stereo Matching with Reliable Disparity Propagation," International Conference on 3D Imaging, Modeling, Processing, Visualization and Transmission, pp. 132-139, 2011.
[10] K.-J. Yoon and S. Kweon, "Adaptive Support-Weight Approach for Correspondence Search," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 28, no. 4, pp. 650-656, 4 2006.
[11] Z. Ke, L. Jiangbo and G. Lafruit, "Cross-based local stereo matching using orthogonal integral images," IEEE Transactions on Circuits and Systems for Video Technology, vol. 19, no. 7, pp. 1073-1079, 2009.
[12] M. Calonder, V. Lepetit, C. Strecha and P. Fua, "BRIEF: Binary Robust Independent Elementary Features," European Conference on Computer Vision, vol. 6314, pp. 778-792, 2010.
[13] K. Zhang, J. Lu and G. Lafruit, "Cross-Based Local Stereo Matching Using Orthogonal Integral Images," IEEE Transactions on Circuits and Systems for Video Technology, vol. 19, no. 7, pp. 1073-1079, 7 2009.
[14] D. Scharstein and R. Szeliski, "A taxonomy and evaluation of dense two-frame stereo correspondence algorithms," IJCV, vol. 47, no. 1-3, pp. 7-42, 2002.
[15] X. Mei, X. Sun, M. Zhou and J. shaohui, "On building an accurate stereo matching system on graphics hardware," IEEE Conference on Computer Vision Workshops (ICCV Workshops), pp. 467 - 474, 2011.
[16] D. G. Lowe, "Distinctive Image Features from Scale-Invariant Keypoints," International Journal of Computer Vision, vol. 60, no. 2, pp. 91-110, 2004.
[17] H. Bay, T. Tuytelaars and L. V. Gool, "Surf: Speeded up robust features," In European Conference on Computer Vision, vol. 3591, pp. 404-417, 2006.
[18] I. E. Sobel, "Camera models and machine perception," Doctoral Dissertation, 1970.
[19] H. Hirschmu¨ ller, "Stereo Processing by Semiglobal Matching and Mutual Information," Pattern Analysis and Machine Intelligence, vol. 2, pp. 328 - 341, 2008.
[20] H. Hirschm¨uller, "Stereo Vision in Structured Environments by Consistent Semi-Global Matching," Computer Vision and Pattern Recognition, vol. 2, pp. 2386 - 2393, 2006.
[21] H. Hirschm¨uller, "Accurate and Efficient Stereo Processing by Semi-Global Matching and Mutual Information," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), vol. 2, pp. 807 - 814, 6 2005.
[22] A. Ansar, A. Castano and L. Matthies, "Enhanced real-time stereo using bilateral filtering," 3D Data Processing, Visualization and Transmission, pp. 455 - 462, 2004.
[23] S. D. and R. Szeliski, "http://vision.middlebury.edu/stereo/eval/," Middlebury stereo evaluation - version 2, 2010.