混合波原子和双边滤波的纹理图像滤波方法

摘要

This paper proposed a novel diffusion scheme by hybridizing wave atoms thresholding for textural images. Also designed a denoising algorithm based on the relationship between solution of nonlinear diffusion equation and bilateral filtering method. Numerical experiments illustrate the good performance in comparison to the state-of-the-art denoising algorithms, such as wave atoms shrinkage method, and the bilateral filtering method, scale mixtures of Gaussians ( GSM) method, non-local mean ( NLM ) method, by using two objective measures: peak signal-to-noise ratio ( PSNR) and structural similarity ( SSIM).%为了更好地去除纹理图像中的噪声,提出了一种新的混合波原子阈值的振荡纹理图像扩散模型；利用扩散方程和图像滤波方法的理论联系,给出了联合双边滤波的图像去噪算法；最后,利用峰值信噪比(PSNR)和结构相似度(SSIM)两个客观图像质量评价指标,与目前流行的图像去噪方法(包括波原子阈值、双边滤波、高斯尺度混合(GSM),以及非局部滤波(NLM))进行比较.实验结果验证了新方法的有效性.