透過您的圖書館登入
IP:3.145.170.67
  • 學位論文

使用顏色漸層向量於SVG漫畫影像壓縮

Using Color Gradient Vector for SVG Comic Image Compression

指導教授 : 張瑞益

摘要


在[1]中,我們提出了一個新的點陣轉換向量圖形的系統。此系統比前人的方法具有更高的壓縮率與更好的圖形品質。然而此系統並不能夠處理在點陣圖形中的顏色漸層。在圖形中的顏色漸層代表從一種顏色沿著向量平滑地轉換到另一種顏色。然而顏色漸層在經過傳統方法向量化後,會有漸層被分割成色塊的情形發生,不但在圖形品質變差也會產生出較大的檔案。因此在本篇論文中,我們提出一個新的顏色漸層向量處理方法,解決顏色漸層在向量圖形中消失的問題並且提供更高的壓縮率以及更好的圖形品質。本論文使用漸層分割與向量搜尋來分割漸層與找出顏色漸層向量。接下來合併不同的漸層區塊與執行後處理程序,將SVG檔案有效地壓縮。最後填入SVG漸層語法來表現出顏色漸層的效果。實驗結果顯示,我們的方法可以同時應用於不同的向量轉換系統,除了可以大幅地減少SVG檔案大小,並提供更好的圖形品質。實際考慮漫畫電子書應用上,我們的轉換方法可以改善在智慧手機上的顯示時間。

並列摘要


In [1], we proposed a method to convert raster images into vector images. This method has higher compression ratio and better image quality than previous method. However, it is not able to process the color gradient in the raster image. Color gradient in the comic image represents the continuously smooth color changes along a vector from one color to another color. In our previous method, after the process of raster-to-vector, the color gradient will be divided into different color regions. It produces not only the poor perceptual quality but also the large files size. Therefore, in this thesis, we propose a new color gradient method to our raster-to-vector system to solve the color gradient problem. Firstly, it applies a color segmentation algorithm to segment raster image as SVG paths and identifies color gradient vector for each path. Then, it merges neighboring paths which have the same gradient vector and executes the post processing to further reduce the file size. Finally, we embed the gradient syntax to show color gradient in SVG. Experimental results show that our method can adapt the different raster-to-vector methods to not only substantially reduce the file size but also provide better image quality. Moreover, we also improve the rendering speed of SVG file on the smart phone.

參考文獻


[1] Ray-I Chang, Y. Yen, and Ting-Yu Hsu, “An XML-based Comic Image Compression,” Lecture Notes in Computer Science, vol. 5353, pp. 563–572, 2008.
[2] Scalable Vector Graphics (SVG), http://www.w3.org/Graphics/SVG/
[3] Extensible Markup Language (XML), http://www.w3.org/XML/
[4] World Wide Web Consortium (W3C), http://www.w3c.org/
[5] S. Battiato, G. Gallo, and G. Messina, “SVG Rendering of Real Images using Data Dependent Triangulation,” In Proceedings of SCCG, pp. 185-192, 2004.

被引用紀錄


Su, C. Y. (2017). 基於機器學習於內容感知SVG漫畫壓縮及其新應用 [doctoral dissertation, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU201703615

延伸閱讀