Title :
Image denoising using learned overcomplete representations
Author :
Sallee, Phil ; Olshausen, B.A.
Author_Institution :
Dept. of Comput. Sci., UC Davis, CA, USA
Abstract :
We describe a method for learning sparse multiscale image representations using a sparse prior distribution over the basis function coefficients. The prior consists of a mixture of a Gaussian and a Dirac delta function, and thus encourages coefficients to have exact zero values. Coefficients for an image are computed by sampling from the resulting posterior distribution with a Gibbs sampler. Denoising using the learned image model is demonstrated for some standard test images, with results that compare favorably with other denoising methods.
Keywords :
Gaussian processes; image denoising; image representation; image sampling; Dirac delta function; Gaussian function; Gibbs sampler; image denoising; learned overcomplete representations; sparse multiscale image representations; sparse prior distribution; Computer science; Distributed computing; Filtering; Image coding; Image denoising; Image representation; Image sampling; Neuroscience; Noise reduction; Testing;
Conference_Titel :
Image Processing, 2003. ICIP 2003. Proceedings. 2003 International Conference on
Print_ISBN :
0-7803-7750-8
DOI :
10.1109/ICIP.2003.1247261