- 作者: 高興同; 王玲玲
- 作者服務機構: 清華大學資訊科學研究所
- 中文摘要: 碎形影像壓縮方法是一種新穎的影像壓縮技術,並且在影像壓縮的領域上,表現出它出色的處理能力。此種影像壓縮方法可產生壓縮率超過一萬比一的情況。在這方法下,一張影像不再以一序列的像素矩陣來表示,而是以一組排列相當緊密的數碼,稱作反複函數系統數碼,來代表影像的各個分割區的資訊,然後再由這些分割區的資訊建構出完整影像。在本篇論文中,我們提出利用N步搜尋方法來加速碎形影像的壓縮速度,並探討在反複函數系統數碼上作影像處理之可行性。當一張影像被壓縮成反複函數系統數碼後,一些影像處理的工作就可以接著進行。這些工作包括雜訊減少,邊界偵尋,邊界之放大與縮小等,都可以在碎形影像壓縮之解壓縮的過程中同時進行。此外,反複函數系統數碼對紋理分割工作而言,是一組良好的特徵值,故可直接用來作紋理分割工作。由論文中所提供實驗結果顯示,我們所提出的方法的確具有可行性。
- 英文摘要: The fiactal image compression scheme is shown to be powerful in image compression In this method,an image is not represented in an array of pixels; instead, a relatively compact set of numbers, callediterated function system (IFS)codes, is used to code the segments that compose the image In this paper,a speed improvement in the fractal compression technique by using the N-step matching algorithm isproposed, and the feasibility of image processing based on IFS codes is presented. This scheme includesnoise reduction, edge detection, and edge magnification or demagnification, which are jointly implementedwith the decompression procedure instead of being treated as separate operations. In addition, texturesegmentation can be achieved by using the IFS codes as features. Experimental results are presented toshow the feasibility of the proposed approach
- 中文關鍵字: fractal image compression; iterated function system; edge derection; texture segmentation
- 英文關鍵字: --