徒手持相機取像時，可能因手振晃動或其它人為因素，使成像有偏斜(存在角差)。在本論文中，我們提出一方法，係利用左右對稱取像標的物(如人體)對稱軸與重力方向同向的事實，希望據以藉由電腦輔助估計一影像在取像時歪斜的程度(旋轉軸在鏡頭中心軸方向情況)，並加以調整改善。

我們所根據的是：(1)左右對稱物體(但存在有角差情況)在該角度之投影值亦會呈現左右對稱情況(除有一位移外)；(2)由中截投影定理(Central Slice Projection Theorem)，原物體於二維富氏轉換後之極座標表示中，其相位(phase)於上述角度處，當對稱物位置及偏斜度符合合理限制下(本論文中有推得相關條件)，在一定範圍內具線性特性。之後，我們找出富氏轉換相位(phase)在(2)中所提上述範圍內最具線性性質之角度（採用Least square fitting方式），並以之為偏斜角度估計值。有了估計的偏斜角度，即可透過反旋轉作修正。

我們針對所提方法，以模擬對稱圖形及實際存在左右對稱物(人)之影像作測試，發現所提方法在對影像加入二值化前處理後相當可行，其中估計偏斜準確度可達1°範圍。除了對稱軸角度外，本論文所提方法可附帶推算出對稱軸所在位置，亦可使用於需找對稱軸之其它影像應用。

[1] R.N. Bracewell, “Strip intergration in radio astronomy,” Aust. J.Phys., 9, 198-217, 1956.

[2] H. Peng, and H. Stark, “Direct fourier reconstruction in Fan-Beam tomography,”, IEEE Trans. on Medical Image, Vol. mi-6, No. 3, September 1987. pp.209-219.

[3] S. Derrode, and F. Ghorbel,“Shape analysis and symmetry detection in gray-level objects using the analytical Fourier–Mellin representation,” Signal Processing, Vol. 84, No. 1, January 2004. pp. 25-39.

[4] B. Rousso, S. Avidan, A. Shashua, and S. Peleg , “Robust recovery of camera rotation from three frames,” Computer Vision and Pattern Recognition, 1996. Proceedings CVPR '96, 1996 IEEE Computer Society Conference on 18-20 June 1996 Page(s):796 – 802.

[5] R.K.K. Yip, “A Hough transform technique for the detection of reflectional symmetry and skew-symmetry,” Pattern Recognition Letters, Vol. 21, No. 2, February 2000. pp. 117-130.

[6] T.J. Cham, and R. Cipolla, “Symmetry detection through local skewed symmetries,” Image and Vision Computing Vol. 13, No. 5, June 1995. pp. 439-450.

[7] Y. Lei, and K.C. Wong, “Detection and localisation of reflectional and rotational symmetry under weak perspective projection,” Pattern Recognition Vol. 32, No. 2, 1999. pp. 167-180.

[8] D. Shen, H.H.S. Ip, and E.K. Teoh , “Robust detection of skewed symmetries by combing local and semi-local affine invariants,” Pattern Recognition, Vol. 34, No. 7, 2001. pp. 1417-1428.

[9] W. Wen, and B. Yuan, “Skewed symmetry detection of closed contours based on their geometric properties,” Computer Architectures for Machine Perception, 1995. Proceedings. CAMP '9518-20 Sept. 1995 Page(s):161 – 168.

[10] J.M. Inesta, M. Buendia, and M.A. Sarti, “Local symmetries of digital contours from their chain codes,” Pattern Recognition Vol. 29, No. 10, 1996. pp. 1737-1749.

[11] C. Sun, “Symmetry detection using gradient information,” Pattern Recognition Letters, Vol. 16, No. 9, September 1995. pp. 987-996.

[12] C. Sun., and D. Si., “Fast reflectional symmetry detection using orientation histograms,”Real-Time Imagine. Vol. 5, No. 1, February 1999. pp. 63-74.

[13] D. Shen, H.H.S. Ip, and E.K. Teoh, “An energy of asymmetry for accurate detection of global reflection axes,” Image and Vision Computing, Vol. 19, No. 5, April 1, 2001. pp. 283-297.

[14] C.E. Shannon,“Communication in the presence of noise,” In Proceedings of IRE, Vol. 38, No. 4, April 1950. pp.417-425.

[15] G. Iannizzotto, P. Lanzafame, and F. L. Rosa, “A CNN-based framework for 2D still-image segmentation,”Computer Architecture for Machine Perception, 2005. CAMP 2005. Seventh International Workshop on Proceedings. 4-6 July 2005 Page(s):210 – 215.

[16] A.K. Jain, Fundamentals of digital image processing, Prentice Hall, 1989.

[17] S.M. Kao, T.K. Ning, C.W. Wu, and H. Peng, “Extraction of the focused object in an image by filtering out the defocused background,” In Proceeding of IEEE 1994 International Symposium on Speech, Image Processing, &Neural Networks, Hong Kong, April 13-16, 1994.