現今隨著大螢幕顯示器的普及，人們對於影像解析度大小的要求越來越高，但是由於數位影像擷取時可能限制於設備的功能或儲存容量的大小，往往只能取得解析度較低之影像，如果要將這些低解析度數位影像轉移於另一個高解析度之設備播放時，就必須運用影像處理中放大的技術，而影像內插就是一種常用於影像放大時的技術，但是傳統影像內插技術存在著許多問題，例如影像放大後會使得物體邊緣模糊或產生鋸齒狀，因而降低影像的視覺品質。

本論文提出一種判斷邊緣方向的方法，將邊緣的資訊區分為五大類，並且根據不同類型的邊緣資訊，設計五種對應的參考像素樣板，最後再提出一種結合影像中像素距離與結構相似性之間的關係，根據像素之間彼此距離的遠近以及像素周圍結構的相似性，為每一個參考的像素定義最合適的權重，經由加權總合內插得到放大後所遺失的像素。經由實驗結果可以顯示出，在影像放大之後，無論是任何角度的邊緣，比起其他方法都可以得到較清晰的視覺品質以及較高的PSNR(Peak Signal to Noise Ratio)值。

關鍵字：影像放大、判斷邊緣方向、參考像素樣板、非線性內插

As large screen display devices become popular, people ask high-resolution images. However, due to the limitation of the digital image cameras, the captured images tend to be the low-resolution images. If we display these low-resolution images on a high-resolution display device, we must use a super-resolution technique of the image processing. The image interpolation is one common technique to do super-resolution. The traditional image interpolation techniques have many problems, such as the blurring edges or jaggies on the edge region, therefore reduces the visual quality of the image.

In our thesis, we proposed a simple edge direction detection method which classifies the local information of a missing pixel into five categories when we want to interpolate it for a high-resolution image. According to its type of edge direction we designed five templates of the reference pixels. Finally, we proposed a nonlinear interpolation algorithm which combines the geometric distance weight and patch similarity weight to determine an optimal weight for every reference pixel. Through the weighted sum of all reference pixels then we interpolate the missing pixels. The experiment result shows that no matter any direction of the edges, the proposed algorithm can obtain sharper edges and clear visual quality than other methods, and it also can increase the PSNR.

Key word: super-resolution, edge direction detection, templates of the reference pixels, nonlinear interpolation

References
[1] T. Lehmann, C. Gönner, and K. Spitzer, “Survey: Interpolation methods in medical image processing,” IEEE Trans. Med. Imag., vol. 18, no. 11, pp. 1049–1075, Nov. 1999.
[2] T.Blu, P. Thevenaz, and M. unser, “Linear interpolation revitalized,” IEEE Trans. Image Processing, vol. 13, no.5, pp.710-719, may 2004.
[3] R. Gonzalez and R. Woods, digital Image Processing, 2nd ed. Upper Saddle River, Nj: Prentice-Hall, 2002.
[4] H.S. Hou and H.C.Andrews,”Cublic splines for image interpolation and digital filtering,” IEEE Trans. Acoust., Speech, Signal Process.,vol. ASSP-26, no.6, pp.508-517,Jun.1978.
[5] T. Lehmann, C. Gönner, and K. Spitzer, “Survey: Interpolation methods in medical image processing,” IEEE Trans. Med. Image., vol. 18, no.11, pp.1049-1075.
[6] R. G. Keys, “Cubic convolution interpolation for digital image processing,” IEEE Trans. Acoust., Speech, Signal Processing, vol. ASSP-29, pp. 1153–1160, Dec. 1981.
[7] Xiaojun Feng and Jan P. Allebach, “segmented image interpolation using edge direction and texture synthesis,” ICIP 2008. 15th IEEE international Conference on Image Processing 2008, pp.881-884.
[8] Yang Yun-feng, Su Zhi-xun, HU Jin-yan,” A New Edge-holding Algorithm of Image Interpolation,” Journal of Image and Graphics, Vol. 10, No. 10, pp. 1248-1251., Oct., 2005.
[9] X. LI and M. Orchard, “New edge-directed interpolation,” IEEE trans. Image Process., vol. 10, pp.1521-1527, Oct. 2001.
[10] S. Park, M. Park, and M. Kang, “Super-resolution image reconstruction: a technical overview,” IEEE Signal Process. Mag., vol. 20, no. 3, pp. 21–36, May. 2003.
[11] R. C. Gonzalez and R. E. Woods, "Digital Image Process," Prentice- Hall, N.J., 2002.
[12] Zhou Dengwen, “An Edge-Directed Bicubic Interpolation Algorithm,” IEEE Trans. Image and Signal Processing (CISP), 2010 3rd International Congress on, 1186 – 1189, 16-18 Oct. 2010.
[13] Fujian Wang, “A New Nonlinear Interpolation Algorithm for Edge Preserving”, IEEE Trans. Multimedia Technology (ICMT), 2010 International Conference on, 1 – 4, 29-31 Oct. 2010
[14] C. Tomasi and R. Manduchi, “Bilateral filtering for gray and color images,” in Proc. Int. Conf. Computer Vision, 1998, pp.839–846.
[15] A. Giachetti, N. Asuni, Fast artifacts-free image interpolation, in Proceedings of the British Machine Vision Conference, Leeds, UK, 2008, pp. 123–132.
[16] Kumar P, Sengupta K, Lee A. A comparative study of different color spaces for foreground and shadow detection for traffic monitoring system. In: Proceedings of the 5th IEEE International Conference on Intelligent Transportation Sys- tems. Singapore, Singapore: IEEE, 2002. 100-105
[17] MathWorks. Image Processing Toolbox User0s Guide. Nat-ick: MathWorks, 2009