臺灣博碩士論文加值系統

在科技的快速進步下，人們對於真實感的需求更加提升，但生活周遭還是充滿了許多低解析度的視訊，而視訊超解析(Super Resolution, SR)能夠將這些低階析度的影像提升到高解析度影像。
現有的攝影器材提供了錄影同時拍照的功能，於是我們提出了一個超解析度的架構，我們利用較高解析度的影像加入到視訊超解析當中，透過影像自我相似性對視訊每個幀(Frame)進行超解析，並使用金字塔架構及二元搜尋法來加速補丁匹配運算。
系統提供使用者將欲超解析的視訊作為輸入，系統將透過影像的自我相似性來補償原視訊不足的部分，以提升視訊之畫質，達到高解析度的輸出結果。

Because of the advances in Science and technology, people are more used to realistic things, but there are still a lot of low-resolution images and videos around the world, so we need the Super Resolution technique to improve the quality.

Nowadays, Hybrid-cameras have been invented, and therefore we propose a new technique that extends existing example-based super-resolution frameworks, by considering the embedded external high definition images for video super resolution, together with the property of local self-similarity. For the efficacy of patch matching, we use a pyramid model and binary search method to find the matching window quickly and reduce comparison time.

1. 緒論 2
1.1. 研究動機與目的 2
1.2. 論文架構 2
2. 相關文獻 3
2.1. 影像超解析 4
2.1.1. 插值法(Interpolating) 4
2.1.2. 學習法(Learning-based) 6
2.2. 視訊超解析 9
2.2.1. 學習法(Learning-based) 9
2.2.2. 移動估計(Motion-based) 13
3. 超解析演算法 15
3.1. 自我相似性 15
3.2. 演算法流程 16
3.3. 非二元濾波器 18
3.4. 補丁匹配演算法 22
3.4.1. 金字塔模型 22
3.4.2. 二元搜尋法 24
3.5. 系統優化 25
3.5.1. 演算法加速 25
3.5.2. 硬體加速 26
4. 實驗分析與結果 27
4.1. 二元搜尋法範圍誤差分析 28
4.2. 超解析演算法效率分析 29
4.3. 邊緣偵測分析 30
4.4. 影像超解析結果 31
4.5. 視訊超解析結果 33
5. 結論與未來展望 36
6. 參考文獻 37

[1.]Xiangjun Z. and Xiaolin W. 2008. Image Interpolation by Adaptive 2-D Autoregressive. IEEE Transactions On Image Processing, Volume 17, pp. 887 – 896.
[2.]Freedman G. and Fattal R. 2011. Image and Video Upscaling from Local Self-Examples. ACM Transactions on Graphics, Volume 30, Issue 2.
[3.]Jianchao Y. Zhaowen W., Zhe L., Scott C. and Thomas H., 2012. Coupled dictionary training for image super-resolution. IEEE Transactions On Image Processing, Volume 21, pp. 3467-3477.
[4.]Basavaraja SV., Ajit S. B. and Sudha V., 2010. Detail Warping Based Video Super-Resolution Using Image Guides. IEEE 17th International Conference on Image Processing, pp.2009-2012.
[5.]Shan, Q., Li, Z., Jia, J., Tang, C. 2008. Fast Image/Video Upsampling. ACM Transactions on Graphics. Graph. 27, 5, Article 153(December 2008), 7 pages. DOI = 10.1145/1409060.1409106 http://doi.acm.org/10.1145/1409060.1409106.
[6.]López S., Callicó G.M., Tobajas F., López J.F. and R. 2009. A Novel Real-Time DSP-Based Video Super-Resolution System. IEEE Transactions on Consumer Electronics, Volume 55, No. 4, pp.2264-2270.
[7.]Glasner, D., Bagon, S., and Irani, M. 2009. Super-resolution from a single image. International Conference on Computer Vision, pp.349–356.
[8.]Fattal, R. 2007. Image upsampling via imposed edge statistics. ACM Transactions on Graphics, Volume 26, Issue 3, No. 95.
[9.]Mallat, S. 1999. A Wavelet Tour of Signal Processing, Second Edition (Wavelet Analysis & Its Applications). Academic Press. ISBN:0123743702 9780123743701.
 
[10.]Freeman W. T., Jones, T. R., and Pasztor, E. C. 2002. Example-based super-resolution. IEEE Comput. Graph. Appl. 22, 2 (March), pp.56–65.
[11.]Brandi, F., de Queiroz R. and Mukherjee D., 2008. Superresolution of video using key-frames and motion estimation, In Proc. IEEE Image Processing (ICIP), pp.321–324.
[12.]Suetake, N., Sakano, M., and Uchino, E. 2008. Image superresolution based on local self-similarity. Journal Optical Review, Volume 15, Issue 1, pp.26-30.
[13.]Yu Zhu, Yanning Zhang and Yuille, A.L., 2014. Single Image Super-resolution using Deformable Patches. IEEE Computer Vision and Pattern Recognition, pp. 2917 – 2924.
[14.]Chen-Chiung Hsieh, Yo-Ping Huang, Yu-Yi Chen and Chiou-Shann Fuh, 2011. Video Super-Resolution by Motion Compensated Iterative Back-Projection Approach. Journal Of Information Science And Engineering 27, pp.1107-1122.
[15.]Ce Liu and Deqing Sun, 2011. A Bayesian Approach to Adaptive Video Super Resolution. IEEE Computer Vision and Pattern Recognition. Pp.209-216.