Theoretical analysis on Noise2Noise using Stein's Unbiased Risk Estimator for Gaussian denoising: Towards unsupervised training with clipped noisy images.
Magauiya ZhussipShakarim SoltanayevSe Young ChunPublished in: CoRR (2019)
Keyphrases
- noisy images
- gaussian noise
- denoising
- noise level
- additive noise
- additive gaussian noise
- noise model
- image denoising
- noise free
- theoretical analysis
- denoising methods
- noise removal
- noise variance
- denoising algorithm
- noise free image
- total variation
- image processing
- natural images
- wavelet denoising
- wavelet shrinkage
- wavelet domain
- image restoration
- edge detection
- noisy data
- impulse noise
- gaussian filter
- wiener filter
- signal to noise ratio
- noise reduction
- total variation minimization
- noise suppression
- higher order
- wavelet packet
- restored image
- training set
- similarity measure
- maximum likelihood
- image patches