摘要
在双树复数小波的基础上引入平移不变的重要思想,使其具有真正意义上的平移不变性.提出了一种基于平移不变双树复数小波变换的图像去噪方法.实验结果表明该方法能有效提高图像去噪效果及图像边缘检测能力.
A novel dual tree complex wavelet thresholding based on translation-invariant theory is introduced to denoise image. Because of the only approximate shift invariant property of dual tree complex wavelet,the method is used to denoise image that still appears light gibbs artifacts. The new method makes it had full shift invariant property. Compared with the transform invariance wavelet method, the new approach can remove the visual artifacts and obtain better visual result and higher PSNR.
出处
《光子学报》
EI
CAS
CSCD
北大核心
2008年第3期604-608,共5页
Acta Photonica Sinica
基金
国家自然科学基金(60375021)
教育部留学回国人员科研启动基金
湖南省杰出青年基金(05JJ10011)
湖南省教育厅重点项目(04A056
05C092)资助
关键词
复数小波
双树复数小波
图像去噪
平移不变
Complex wavelet
Dual tree complex wavelet
Image de-noising
Transform invariance
