改进分形模型的图像降噪算法

改进分形模型的图像降噪算法 Image Denoise Based on Modified Fractal Model 李吉星 LI Ji-xing 1 2 first-author 刘小奇 LIU Xiao-qi 1 2 魏国强 WEI Guo-qiang 1 2 武汉大学电子信息学院 武汉大学电子信息学院 School of Electronic Information, Wuhan University School of Electronic Information, Wuhan University 武汉大学电子信息学院,湖北,武汉,430072 武汉大学电子信息学院,湖北,武汉,430072 利用分形几何中的迭代函数系统理论,针对空间域建立图像分形模型存在相似匹配速度慢的缺点,提出了一种变换域的改进分形模型方法,该方法用频域的搜索匹配代替空间域的搜索匹配,不但域块和区块达到相似匹配的速度更快,而且可以自适应的调整参与计算相似匹配的参数个数.实验结果表明改进的分形模型应用于图像降噪不但可以有效的保护图像的细节和边缘信息,并避免原方法在降噪的同时带来块效应的缺点. Based on the principle of Iterated Functions Systems of Fractal, and in order to filter Gaussian noise in the degraded image and preserve, at the same time, the details of image, in this paper, we advance a improved denoising algorithm based on fractal model. Its characteristics are shown as following:Replacing spatial seaching & matching algorithm by s & m algorithm in frequency domain, it has much quicker speed of matching and convergence; It can auto adjust parameters while in doing matching. It is shown that the new scheme can preserve the details of the image effectively and reduce the block effect. 分形模型 迭代函数系统 仿射变换 压缩映射 fractal model IFS（iterated functions system） affine transformation compress mapping TN911.73 国家自然科学青年基金资助项目（40204008） 2004-01-01 2021-04-01 1