Contourlet收缩和全变差相结合的图像去噪

Image denoising combining Contourlet shrinkage with total variation

充分保持细节的图像去噪在图像处理领域具有重要的意义。一种新的将Contourlet收缩和全变差相结合的混合去噪算法被提出。利用空域自适应的全变差,对含噪图像与Contourlet硬阈值收缩图像的差值图像进行滤波。再和收缩图像相叠加,从而得到最终的去噪图像。实验结果表明,和现有的典型去噪方法相比较,所提出的算法在有效去除噪声和Gibbs伪影的同时,更好地保持了边缘和纹理等重要的细节信息。

Image denoising with details sufficiently preserved has important significance in image processing field.A novel hybrid denoising algorithm is proposed,which combines the contourlet shrinkage method with the total variation approach.Using the spatially adaptive total variation,the difference image between the noisy image and the contourlet hard-threshold shrinkage image is filtered.The filtered difference image is added back to the shrinkage image to get the final denoised image.Experimental results show that,compared with existing typical denoising methods,the proposed algorithm can effectively remove the noise and Gibbs-like artifacts,and has superior performance in preserving the important detail information,such as edges and textures.