一种基于Curvelet变换的图像去噪方法

A Novel Image Denoising Method Based on Curvelet Transform

提出了一种Curvelet变换与循环平移相结合的图像去噪方法,消除了由于Curvelet变换缺乏平移不变性而产生的图像失真.实验结果表明,用该方法去噪后的图像与传统小波去噪图像相比,PSNR值更高,视觉效果也更好.

A novel image denoising algorithm is proposed, which utilizes curvelet transform combining with cycle spinning. It avoids image artifacts due to the lack of translation invariance of curvelet transform. The experimental results show that the proposed curvelet reconstructions method yields denoised images with higher PSNR and better perceptual quality than the ones denoised by classical wavelet transform.