Curvelet Based Denoising Scheme for Recursive Image Debluring
Ippr Lab · Signal Processing · 2005
An image restored by a recursive solution of traveling wave equation is natural to visual interpretation, without ringing effect and distortion emerging, but accumulated noise appears on it obviously. Curvelet transform has variable anisotropy, its projections on different directions assure the noise averaging characteristics. Experimental results presented in this paper suggest that simple thresholding of the curvelet coefficients exhibits higher denoising quality.