一种基于角度的正交FRIT变换及其在图像去噪中的应用

M.N.Do和M.Vetterli提出的Finite Ridgelet Transform(FRIT)因其对线奇异特征的高效表示能力而被广泛应用,但其在图像压缩、去噪的处理中却受到"环绕"现象的严重影响.本文在揭示"环绕"现象和FRAT域系数关系的基础上,提出一种基于角度的正交FRIT变换方案(Angle-based FRIT,AFRIT).该方案具有更好的能量集中特性,并有效地降低了"环绕"现象.进一步,对正交FRIT去噪问题建模,并提出了一种基于AFRIT的改进阈值.不同噪声水平下各种图像的去噪实验结果表明,采用改进阈值的AFRIT同现在常用的FRIT去噪方法相比,具有明显的优越性.

Angle Based Orthogonal FRIT and Its Application in Image Denoising

Finite Ridgelet Transform(FRIT) is a widely used orthogonal transform for its good property of effectively representation of linear singularities in image.But the limitation of "wrap around" effect restricts its applications in image denoising and compression.In this paper,based on analysis of the relation between "wrap around" effect and distribution of FRAT(Finite Radon Transform) coefficients,we proposed an angle-based orthogonal FRIT scheme(AFRIT) which can concentrate the coefficients energy and reduce the influence of "wrap around" more greatly than the usual ones.We then modeled the denoising problem and proposed a novel threshold selecting method based on AFRIT.Experiments carried out on different images with varying levels of addictive white Gaussian noise show AFRIT's validity.