机构地区: 中山大学数学与计算科学学院科学计算与计算机应用系
出 处: 《计算机科学》 2006年第9期222-224,共3页
摘 要: 通过迭代函数系统(IFS)的不动点来逼近源图像的分形图像压缩方法是图像编码的一种相对新的技术。目前这种方法已派生出众多的图像编码方案,其中大多采用分块和匹配的方法来实现对图像的编码。为提高计算效率,总是希望能用尽可能少的城块(Domain Block)为图像的分类块(Range Block)找到最佳匹配。但这种考虑容易导致最终获得一个有些粗糙的图像编码。本文提出了一类预处理-修正模式的分形图像编码方法。我们保留原有编码作为预编码,进而提出修正预编码的具体算法。算法中充分利用了已有的计算结果,且修正编码过程中可以适当地加入人工干预,有利于提高压缩效率和改进编码质量。 Fractal image compression is a relatively new technique for encoding images compactly through constructing an iterative function system (IFS) whose fixed point will approximate the original image. Now this broad principle has encompassed a very wide variety of coding schemes, many of which encode digital image based on the block dividing and block-matching. For improving computation efficiency, one always hope to find the best matching domain block in a range which is as small as possible. However, this kind of consideration is apt to result in getting a coarse codes finally. This paper proposes a fractal Image compression method based on pretreatment-modification mode, which keeps the coarse image codes as a pre-codes, and then presents a modification algorithm to improve on the original codes. In the scheme, the compression efficiency and coding quality can be improved on by utilizing the existing result and Adopting certain manual intervention.
领 域: [自动化与计算机技术] [自动化与计算机技术]