【摘要】 综合利用Contourlet变换和递归Cycle Spinning,提出一种新的图像去噪方法.由于Contourlet变换缺乏平移不变性,直接进行Contourlet系数阈值图像去噪会产生伪吉布斯现象(导致图像失真),本文引入递归Cycle Spinning来有效地消除这种由于Contourlet变换缺乏平移不变性而产生的图像失真.实验结果显示,与小波递归Cycle Spinning图像去噪等方法相比,该方法明显改善图像视觉效果,显著提高图像的PSNR值.更多还原

【Abstract】 A new method for image de-noising which colligated the strongpoint of Contourlet transform and Recursive Cycle Spinning was presented.Due to the lack of translation invariance of the Contourlet transform,image de-noising by coefficient thresholding would lead to Gibbs-like phenomena(lead to artifacts).Recursive Cycle Spinning was employed to avoid the artifacts.The experimental results indicate that the method can get better visual effect and PSNR value compared with the methods like wavelet image de-noising using the Recursive Cycle Spinning.更多还原