影像超解析技術是從一張低解析度的影像生成一張高解析度的影像並有著良好的視覺品質。在本論文中，我們提出一個以向量量化為基底的超解析演算法，而建立一個良好的編碼簿是能決定以向量量化為基底的超解析演算法是否能夠產生高視覺品質的一個重要因素，因此本研究為使「生成編碼簿」以及後續「在重建影像時使用編碼簿」時有其一致性，在「生成編碼簿」時改以用相關係數作為分群方式。此外，因新式編碼簿中的編碼向量為單位向量，在重建影像時，將以正規化參數乘以被選中的編碼向量。由實驗結果顯示，本研究所提出的方式確實能增進影像的視覺品質並提供自然的細節。

Super resolution technology is generating a high-resolution (HR) image from a low-resolution (LR) image to get a better visual quality. In this thesis, a vector quantization (VQ) based super resolution algorithm is proposed. Training a proper codebook is one of the significant factors in VQ. To let the distance metric in training codebooks consistent with that in image reconstruction procedure, the proposed method utilizes correlation coefficient as a new distance metric to generate a new kind of codebooks. Since the new codebooks contain unit vectors, the normalizing parameter will be multiplied by the selected code vectors in image reconstruction procedure. Experimental results show that the proposed algorithm truly improve the visual quality with “natural detail”.

[1] H. Hou and H. Andrews, “Cubic splines for image interpolation and digital filtering,” in IEEE Transactions on Acoustics, Speech and Signal Processing, vol. 26, no. 6, pp. 508 – 517, Dec. 1978.
[2] C. E. Duchon, “Lanczos filtering in one and two dimensions,” in J. Appl. Meteor., vol. 18, no. 8, pp. 1016 – 1022, Aug. 1979.
[3] S. Dai, M. Han, Y. Wu, and Y. Gong, “Bilateral back-projection for single image super resolution,” in IEEE International Conference on Multimedia and Expo, pp.1039 – 1042, July 2007.
[4] W. Dong, L. Zhang, G. Shi, and X. Wu, “Nonlocal back-projection for adaptive image enlargement,” in IEEE International Conference on Image Processing, pp. 349 – 352, Nov. 2009.
[5] H. Ji and C. Fermuller, “Robust wavelet-based super-resolution reconstruction: theory and algorithm,” in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, no. 4, pp. 649 – 660, April 2009.
[6] S. Mallat and G. Yu, “Super-resolution with sparse mixing estimators,” in IEEE Transactions on Image Processing, vol.19, no. 11, pp. 2889 – 2900, Nov. 2010.
[7] Y. C. Wee and H. J. Shin, “A novel fast fractal super resolution technique,” in IEEE Transactions on Consumer Electronics, vol. 56, no. 3, pp. 1537 – 1541, Aug. 2010.

[8] W. T. Freeman and E. C. Pasztor, “Learning low-level vision,” in IEEE International Conference on Computer Vision, vol. 2, pp. 1182 – 1189, 1999.
[9] H. Chang, D.-Y. Yeung, and Y. Xiong, “Super-resolution through neighbor embedding,” in IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 1, pp. I, June – July 2004.
[10] S.-H. Hong, R.-H. Park, S. Yang, and J.-Y. Kim, “Image interpolation using interpolative classified vector quantization,” in Image and Vision Computing, vol. 26, no.2, pp.228 – 239, 2008.
[11] T.-M. Kuo and S.-C. Tai, “Super-resolution algorithm based on weighted vector quantization,” in Journal of Electronic Imaging, vol. 23, Issue 2, March 2014.
[12] H.-M. Hang and B. Haskell, “Interpolative vector quantization of color images,” in IEEE Transactions on Communications, vol. 36, pp. 468 – 470, April 1988.
[13] Z. Wang, A. Bovik, H. Sheikh, and E. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” in IEEE Transactions on Image Processing, vol.13, no. 4, pp. 600 – 612, April 2004.
[14] X. Li, K. M. Lam, G. Qiu, L. Shen and S. Wang, “Example-based image super-resolution with class-specific predictors,” in Journal of Visual Communication and Image Representation, vol.20, no. 5, pp. 312 – 322, July 2009.