CAJ | 학술논문

本文提出了一种基于非下采样Contourlet变换与非线性各向异性扩散的方法进行含噪图像的去噪和增强。首先对含噪图像进行非下采样Contourlet分解，对每个分解层的各个子带进行非线性收缩和拉伸，以达到抑制噪声和增强图像特征的目的。然后，对去噪增强后图像的Contourlet小系数进行空间域的非线性各向异性扩散，以去除由于进行非下采样Contourlet去噪所造成的为伪Gibbs现象和 side-band效应。实验结果表明，本文方法相比于无扩散的Wavelet和Contourlet方法相比，不仅对图像进行了去噪和增强，而且有效的抑制了伪Gibbs现象和 side-band效应。
본문제출료일충기우비하채양Contourlet변환여비선성각향이성확산적방법진행함조도상적거조화증강。수선대함조도상진행비하채양Contourlet분해，대매개분해층적각개자대진행비선성수축화랍신，이체도억제조성화증강도상특정적목적。연후，대거조증강후도상적Contourlet소계수진행공간역적비선성각향이성확산，이거제유우진행비하채양Contourlet거조소조성적위위Gibbs현상화 side-band효응。실험결과표명，본문방법상비우무확산적Wavelet화Contourlet방법상비，불부대도상진행료거조화증강，이차유효적억제료위Gibbs현상화 side-band효응。
This paper proposes a hybrid method for image enhancement and noise reduction by using the NonSubsampled Contourlet Transform (NSCT) and Nonlinear Anisotropic diffusion. For the purpose of reducing pseudo-Gibbs and Contourlet-like artifacts, an improved gain function which integrates feature enhancement and noise reduction is introduced to nonlinearly shrink and stretch the NSCT coefficients firstly. Then, the enhanced results are further processed by the nonlinear diffusion where only the nonsignificant, i.e., thresholded, NSCT coefficients are changed by means of a diffusion process in order to reduce pseudo-Gibbs artifacts. Numerical experiments show its good performance to cancel oscillations near the discontinuities and eliminate Contourlet-like artifacts while preserving the strong edges and shapes of the features of surfaces, in comparison to existing methods.