基于Curvelet变换的软硬阈值折衷图像去噪

与小波变换相比,Curvelet变换更好地表达图像的边缘和细节,因此更适合多尺度图像去噪。针对软阈值和硬阈值去噪方法存在的不足,提出了基于Curvelet变换域的软硬阈值折衷去噪法,并采用不同的阈值自适应地对不同的Curvelet子带进行阈值化。实验结果表明该方法对图像中的边缘、弱的直线和曲线特征有更好的恢复。去噪后图像PSNR值更高,视觉效果更好。

A image denoising method between soft and hard thresholding based on Curvelet transform

The curvelet transform represents edges and details better than wavelets,and is therefore well-suited for multiscale image denoising.According to the defects of soft thresholding and hard thresholding image denoising methods,the method between soft and hard thresholding image denoising approach in curvelet domain is proposed,and the curvelet transform coefficients in different subbands are filtered with adaptive thresholds.Experiment results show that the new method yields denoised images with higher quality recovery of edges and of faint linear and curvilinear features.It is capable of achieving the higher peak signal-to-noise ratio(PSNR) and giving better visual quality.