A principled approach to image denoising with similarity kernels involving patches.
Arnaud de DeckerJohn Aldo LeeMichel VerleysenPublished in: Neurocomputing (2010)
Keyphrases
- image denoising
- natural images
- denoising
- noisy images
- image patches
- partial differential equations
- total variation
- image processing
- multiscale
- wavelet domain
- denoising algorithm
- wavelet coefficients
- contourlet transform
- higher order
- similarity measure
- object recognition
- translation invariant
- kernel methods
- undecimated wavelet transform
- additive gaussian noise
- denoising methods
- bilateral filtering
- nonlinear diffusion
- image inpainting
- wavelet domain hidden markov tree
- total variation regularization
- nonlocal means
- computational complexity
- image deblurring
- noise removal
- edge preservation
- feature space
- feature vectors
- kernel function
- frequency domain