基于改进Normal Shrink阈值估计的曲波图像降噪

Improved Normal Shrink Denoising Algorithm Based on Fast Discrete Curvelet Transform

结合Curvelet变换的特点改进Normal Shrink阈值算法的尺度参数,提出一种改进的自适应阈值降噪方法。该方法发挥了Curvelet变换对曲线边缘的稀疏表示的特性优点。实验结果表明在图像降噪方面与传统的小波收缩阈值方法相比不但有更好的视觉效果,而且峰值信噪比值也更高。

In this paper, an improved adaptive threshold denoising method based on curvelet transform is proposed. In particular, the scale parameter of Normal Shrink is modified to adapt to the characters of eurvelet transform. As a resuit, algorithm takes the advantage of the features that curvelet represent the curves and edges sparely. Experimental resuits indicate that the proposed method can achieve better visual quality and higher PSNR than traditional shrinking threshold methods of wavelet.