立體匹配演算法能將兩張圖的每個像素匹配，藉由匹配結果獲得視差圖，並利用視差圖獲得的深度資訊應用在三維重建。然而，視差圖的精確度會直接的影響深度資訊，尤其在深度值不連續區域、被物體遮擋區域容易產生錯誤的匹配。
立體匹配演算法主要分為全域、區域兩種方法，而大部分的區域演算法由四個步驟所組成，即計算初始成本、成本聚合、計算視差值、細化。本論文於計算成本步驟中，提出一基於小波的方法並與Census Transform演算法結合，用來加強深度不連續區域的成本計算，而細化步驟中，除了使用傳統的左右一致性偵測異常值，我們使用超像素分割技術切割整張原圖，偵測每個超像素中的異常值進行消除。
實驗結果顯示，所提出基於小波的方法與Census Transform結合和Absolute differences與Census Transform結合的方法相比，平均能降低0.33% 深度值不連續區域之錯誤百分比，而結合的方法相較基於小波的方法與Census Transform分別降低1.8375% 、0.685% 之錯誤百分比，更顯示結合兩方法的必要性。同時，探討細化步驟之性能，結果顯示基於分割之視差圖細化方法能減少區域內的錯誤，使所產生的三維重建較為平滑。

Stereo matching algorithm finds correspondences of pixels in two image to generate disparity maps. Depth information obtained by disparity map can be used to 3D reconstruction. However, accuracy of disparity map directly affect depth information, and errors normally occur in depth discontinuous regions an occluded regions.
Stereo matching algorithm mainly divides into two classes, local algorithm and global algorithm. Most local algorithm consists of four steps, initial cost computation, cost aggregation, disparity computation, and refinement. In our thesis, we propose a wavelet-based method combined with the census transform algorithm in the cost computation step to enhance the accuracy at regions near depth discontinuities. In the refinement step, following the tradition Left-Right Consistency check, we classify each pixel by using superpixel segmentation, detect and correct outliers in superpixels.
The Experiments show that our proposed wavelet-based method combined with the census transform can averagely reduce absolute differences combined with the census transform error percentage of depth discontinuous regions by 0.33%. The combination is necessary because the combination method over the proposed wavelet-based method and the census transform are 1.8375% and 0.685%, respectively. Besides, we discuss the performance of the refinement step. The experiments show smoothness in 3D reconstruction by using the disparity maps generated with the segmentation refinement.

[1] R. Achanta, A. Shaji, Kevin Smith, A. Lucchi, P. Fua, and S. Su¨sstrunk, “SLIC Superpixels Compared to State-of-the-Art Superpixel Methods,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol.34, no.11, 2012.
[2] M. Bleyer and M. Gelautz, “Graph-based surface reconstruction from stereo pairs using image segmentation,” in SPIE, pages vol. 5665: 288–299, 2005.
[3] M. Z. Brown, D. Burschka, , G. D. Hager, “Advances in computational stereo”, IEEE Trans. Pattern Anal. Mach. Intell., vol.25, no.8, pp. 993–1008 ,2003.
[4] P. Compañ, R. Satorre, R. Rizo, “Disparity estimation in stereoscopic vision by simulated annealing,” Artificial Intelligence research and development, IOS Press, Amsterdam, pp. 160–167, 2003.
[5] I. Ernst, H. Hirschmüller, “Mutual information based semi-global stereo matching on the GPU,” in Proceedings of the 4th International Symposium on Advances in Visual Computing, in ISVC 08, Springer-Verlag, Berlin, Heidelberg, pp. 228–239, 2008.
[6] P. F. Felzenszwalb and D. P. Huttenlocher, “Efficient belief propagation for early vision,” in CVPR, pages I: 261–268, 2004.
[7] A. Fusiello, V. Roberto, E. Trucco, “Symmetric stereo with multiple windowing”, Int. J. Pattern Recognit. Artif. Intell., vol.14, no.8, pp. 1053–1066, 2001.
[8] H. Hirschmüller, “Stereo processing by semiglobal matching and mutual information,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 30, no. 2, pp. 328–341, 2008.
[9] G. Hong, B. Kim, “A local stereo matching algorithm based on weighted guided image filtering for improving the generation of depth range images”, Displays, vol.49, pp.80-87, 2017.
[10] X. Huang, C. Yuan, and J. Zhang, “Graph Cuts Stereo Matching Based on Patch-Match and Ground Control Points Constraint,” Springer International Publishing, 2015.
[11] M. Humenberger, C. Zinner, M. Weber, W. Kubinger, and M. Vincze, “A fast stereo matching algorithm suitable for embedded real-time systems,” CVIU 2010, vol.114, no.11, pp.1180-1202, 2010.
[12] T. Kanade, and M. Okutomi, “A stereo matching algorithm with and adaptivewindow: theory and experiments”, IEEE Trans. Pattern Anal. Mach. Intell., vol.16, no.9, 1994.
[13] D. Keen, “A review of 3d reconstruction from video sequences,” in Intelligent Sensory Information Systems Technical Report, University of Amsterdam, 2005.
[14] A. Klaus, M. Sormann, and K. Karner, “Segment-based stereo matching using belief propagation and a self-adapting dissimilarity measure,” in Proc. of International Conference on Pattern Recognition, vol. 3, pp. 15–18, 2006.
[15] J. Kowalczuk, E. T. Psota, L. C. Pérez, “Real-time stereo matching on CUDA using an iterative refine-ment method for adaptive support-weight correspondences”, IEEE Trans. Circuits Syst. Video, vol.23, no.1, pp. 94-104, 2013.
[16] S. Kumar, N. Sukavanam, B. Raman and S. Kumar, “Human Visual System and Wavelet Transform based Disparity Estimation,” Proceedings of 4th International Conference on Emerging Trends in Engineering & Technology (IEEE-ICETET’11), pp. 195-199, Nov 18-20, 2011.
[17] D. G. Lowe, “Object recognition from local scale-invariant features,” in International Conference on Computer Vision, 2, 1999.
[18] X. Mei, X. Sun, M. Zhou, S. Jiao, H. Wang, and X. Zhang, “On building an accurate stereo matching system on graphics hardware,” in Proc. IEEE Int. Conf. Comput. Vis. Workshop, pp. 467–474, 2011.
[19] J. Menant, G. Gautier, M. Pressigout, L. Morin, and J. Nezan, “An automatized method to parameterize embedded stereo matching algorithms,” Journal of Systems Architecture, vol. 80, pp. 92-103, 2017.
[20] C. Rhemann, A. Hosni, M. Bleyer, C. Rother, and M. Gelautz, “Fast cost-volume filtering for visual correspondence and beyond,” IEEE Transactions on Pattern Analysis and Machine Intelligence, col.35, no.2, 2013.
[21] B. Salehian and A. M. Fotouhi, “Dynamic programming-based dense stereo matching improvement using an efficient search space reduction technique,” in Optik, vol. 160, pp.1-12, 2018.
[22] M. Samadi and M. Othman., “A new fast and robust stereo matching algorithm for robotic systems,” IC2IT2013, vol.209, pp.281-290, 2013.
[23] I. Sarkar and M. Bansal, “A wavelet-based multiresolution approach to solve the stereo correspondence problem using mutual information,” IEEE Trans. Syst., Man, Cybern. B, Cybern., vol. 37, no. 4, pp. 1009–1014, 2007.
[24] S. D. Scharstein, R. Szeliski, “A Taxonomy and Evaluation of Dense Two-frame Stereo correspondence Algorithms,” International Journal of Computer Vision , vol. 47, no. 1, pp. 7-42, 2002.
[25] S. D. Scharstein and R. Szeliski. Middlebury stereo vision. [Online]. Available: http://vision.middlebury.edu/stereo
[26] C. Stentoumis, L. Grammatikopoulos, I. Kalisperakis, E. Petsa, and G. Karras, “A local adaptive approach for dense stereo matching in architectural scene reconstruction,” in Proc. 5th Int. Workshop 3D-ARCH, pp.1–8, 2013.
[27] X. Sun, X. Mei, S. Jiao, M. Zhou, and H. Wang, “Stereo matching with reliable disparity propagation,” in Proc. IEEE Int. Conf. 3DIMPVT, pp. 132–139, 2011.
[28] F. Tombari, S. Mattoccia, L. D. Stefano, E. Addimanda, “Classification and evaluation of cost aggregation methods for stereo correspondence”, in CVPR 2008, pp. 1–8, 2008.
[29] C. Wei, S. Lai, and C. Chiang, “Video synthesis from stereo videos with iterative depth refinement,” in Journal of Visual Communication and Image Representation, vol. 47, pp.48-61, 2017.
[30] O. Veksler, “Fast variable window for stereo correspondence using integral images”, Proceedings of the 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol.1, pp. I-556–I-561, 2003.
[31] K. J. Yoon, I. S. Kweon, “Adaptive support-weight approach for correspondence search”, IEEE Trans. Pattern Anal. Mach. Intell., vol.28, no.4, pp. 650–656, 2006.
[32] W. Yu, T. Chen, and J. C. Hoe, “Real time stereo vision using exponential step cost aggregation on GPU,” in Proc. 16th IEEE ICIP, pp. 4281–4284, 2009.
[33] R. Zabih and J. Woodfill, “Non-parametric local transforms for computing visual correspondence,” in Proc. 3rd Int. Eur. Conf. Comput. Vis., pp. 151–158, 1994.
[34] C. Zach, M. Sormann, and K. Karner. “Scanline optimization for stereo on graphics hardware,” in International Symposium on 3D Data Processing, Visualization and Transmission (3DPVT), 2006.
[35] Y. Zhan, Y. Gu, K. Huang, C. Zhang, and K. Hu, “Accurate image-guided stereo matching with efficient matching cost and disparity refinement,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 26, no. 9, pp. 1632–1645, 2016.
[36] K. Zheng, 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, 2009.
[37] Z. J. Zhu, Q. Y. Dai, “Hybrid scheme for accurate stereo matching,” Neurocomputing, vol.252, no.23, pp.24–33, 2017.