這篇論文提出了一個可調性的銳利度強化演算法。本方法可作為一個自動的後處理區塊，套用到自然影像和視訊上，以增進視覺感受並且突顯圖形細節。提出的演算法，包括兩個階段：銳度強化階段與直方圖強化階段。在銳度強化階段，演算法先使用非參考準則 (No-Reference metric) 來偵測影像/視訊全體的銳度，然後再根據此偵測值進行銳度強化。使用的銳化濾鏡並整合人眼視覺系統 (Human Visual System, HVS) 概念。此階段調整不同的影像源，讓它們的銳度更趨均一。在直方圖強化階段，則利用一基於平均亮度的加權式直方圖等化，調整全體直方圖，以突顯原圖的細節對比。其中，透過動態指派高權重予發生機率較小的灰度條，並加入原圖平均亮度為強化條件，可消除直方圖等化時經常出現的過度曝光、亮度偏移等現象，並使亮度變化更為穩定，這在強化視訊時是有助益的。本階段也引進了「對比微調」的區塊，它參考原圖以回復亮暗區在某些情況的強化中失去對比的現象。實驗結果證明，我們的可調性方法能多方適應各種測試影像，在視覺上獲得較佳的強化效果，優於以往的銳化演算法。

Sharpness enhancement improves perceptive experience to digital images. Digital image sources have different kind of blurriness property because of record devices, lossy compression, or upscaling process, thus affecting our visual experience from one image to another. In this thesis, we propose ad adaptive sharpness and enhancement algorithm as an automatic post-process stage for natural image and video. The proposed algorithm intends to increase perspective experience by adjusting acutance and contrast while keeping the original average luminance. The proposed approach consists of two stages: sharpness and histogram enhancement stage. First, In the sharpness stage, overall acutance of an image/video will be detected through a no-reference sharpness metric, and then, a sharpness enhancement filter will be applied based on this acutance metric. The sharpness filter is tried integrating the concept of Human Visual System (HVS) for better perceptual results. The acutance of processed image from different sources become more consistent according to this stage. Second, In the histogram stage, it adjusts the global histogram by a mean-intensity based weighted histogram equalization, which exposes more image detail on originals after enhancing. This is done by a histogram-based weighting mechanic, which not only emphasizes less probability gray-level bins but consider original mean intensity to eliminate the washed-out and intensity-shift phenomenon that often occurs in histogram equalization process. More stable result in mean-intensity variation can be obtained. This is useful to video enhancement. It also introduces a ``contrast refinement' block, which references original pictures to recover the detail lose that happens on brighter and darker region after enhancement.

Experiment results indicate our adaptive algorithm is flexible to several different test sequences, obtaining perceptually satisfied results, and outperforming previous sharpness enhancement algorithms.

[1] G. Arce. A general weighted median filter structure admitting negativeweights. Signal Processing, IEEE Transactions on, 46(12):3195–3205, Dec 1998.

[2] T. Aysal and K. Barner. Quadratic weighted median filters for edge enhancement of noisy images. IEEE Transactions on Image Processing, 2003.

[3] J. Caviedes and S. Gurbuz. No-reference sharpness metric based on local edge kurtosis. Image Processing. 2002. Proceedings. 2002 International Conference on, 3:III–53–III–56 vol.3, 2002.

[4] J. E. Caviedes. Intelligent sharpness enhancement for video post-processing. 2006.

[5] R. Ferzli and L. Karam. A no-reference objective image sharpness metric based on the notion of just noticeable blur (jnb). Image Processing, IEEE Transactions on, 18(4):717–728, April 2009.

[6] M. Fischer, J. Paredes, and G. Arce. Weighted median image sharpeners for the world wide web. Image Processing, IEEE Transactions on, 11(7):717–727, Jul 2002.

[7] R. C. Gonzalez and R. E. Woods. Digital Image Processing (3rd Edition). Prentice-Hall, Inc., Upper Saddle River, NJ, USA, 2006. Sect. 2.1.

[8] W. Ibrahim Ali. Real time video sharpness enhancement by wavelet-based luminance transient improvement. Signal Processing and Its Applications, 2007. ISSPA 2007. 9th International Symposium on, pages 1–4, Feb. 2007.

[9] S.-C. T. Jen-Hao Cheng. A novel high dynamic range compression algorithm with detail refinement. Master’s thesis, National Cheng Kung University, Tainan, Taiwan, R.O.C., 2005.

[10] G. M. Johnson and M. D. Fairchild. Sharpness rules. 2008.

[11] P. Marziliano, F. Dufaux, S. Winkler, and T. Ebrahimi. Perceptual Blur and Ringing Metrics: Application to JPEG2000. Signal Processing : Image Communication, 19(2):163–172, 2004.

[12] S.-C. T. Nai-Ching Wang. A method for automatic image contrast enhancement based on intensity-pair distribution. Master’s thesis, National Cheng Kung University, Tainan, Taiwan, R.O.C., 2008.

[13] R. Pan and X. Meng. A method of local enhancement based on fuzzy set theory. Intelligent Control and Automation, 2000. Proceedings of the 3rd World Congress on, 3:1751–1753 vol.3, 2000.

[14] P. Rieder and G. Scheffler. New concepts on denoising and sharpening of video signals. Consumer Electronics, IEEE Transactions on, 47(3):666–671, Aug 2001.

[15] D. Shaked and I. Tastl. Sharpness measure: towards automatic image enhancement. Image Processing, 2005. ICIP 2005. IEEE International Conference on, 1:I–937–40, Sept. 2005.

[16] A. Taguchi, T. Kimura, and M. Meguro. Sharpening a noisy image by using fuzzy rules. Nonlinear Image Processing IX, 3304(1):144–152, 1998.

[17] J. A. P. Tegenbosch, P. M. Hofman, and M. K. Bosma. Improving nonlinear up-scaling by adapting to the local edge orientation. Visual Communications and Image Processing 2004, 5308(1):1181–1190, 2004.

[18] S. Thurnhofer and S. Mitra. A general framework for quadratic volterra filters for edge enhancement. Image Processing, IEEE Transactions on, 5(6):950–963, Jun 1996.

[19] Q. Wang and R. Ward. Fast image/video contrast enhancement based on weighted thresholded histogram equalization. Consumer Electronics, IEEE Transactions on, 53(2):757–764, May 2007.

[20] P. Whittle. Increments and decrements: Luminance discrimination. Vision Research, 26(10):1677 – 1691, 1986.

[21] M. Wirth and D. Nikitenko. Applications of fuzzy morphology to contrast enhancement. Fuzzy Information Processing Society, 2005. NAFIPS 2005. Annual Meeting of the North American, pages 355–360, June 2005.

[22] L. Yin, R. Yang, M. Gabbouj, and Y. Neuvo. Weighted median filters: a tutorial. Circuits and Systems II: Analog and Digital Signal Processing, IEEE Transactions on, 43(3):157–192, Mar 1996.

[23] S. Zhen-gang, G. Li-qun, and W. Kun. A novel approach to image enhancement and thresholding based on fuzzy theory. Industrial Electronics and Applications, 2007. ICIEA 2007. 2nd IEEE Conference on, pages 2201–2205, May 2007.