Title :
Anisotropic wavelet thresholding for Bayesian image denoising
Author :
Jacovitti, Giovanni ; Neri, Alessandro
Author_Institution :
INFOCOM Dept., Univ. of Rome La Sapienza, Rome, Italy
Abstract :
A new technique for noise suppression in images by wavelet thresholding is presented. The technique focuses on perceptual relevant features, using a multiscale edge oriented wavelet representation. A orientation-dependent zero-memory non-linear soft thresholding rule is defined in the framework of Bayesian MMSE estimation.
Keywords :
Bayes methods; image denoising; image segmentation; least mean squares methods; wavelet transforms; Bayesian image denoising; MMSE estimation; anisotropic wavelet thresholding; multiscale edge oriented wavelet representation; noise suppression; orientation-dependent zero-memory nonlinear soft thresholding; perceptual relevant features; Cleaning; Image coding; Image edge detection; Image resolution; Noise; Noise reduction; Bayesian estimation; Denoising; shrinking; wavelet thresholding;
Conference_Titel :
Signal Processing Conference, 2002 11th European
Conference_Location :
Toulouse
