Estimating hyperparameters of mixture prior using hypothesis-testing problem and its applications to Bayesian image denoising.
Il Kyu EomYoo Shin KimDo Hoon LeePublished in: J. Electronic Imaging (2007)
Keyphrases
- image denoising
- hypothesis testing
- hyperparameters
- posterior distribution
- maximum a posteriori
- bayesian inference
- prior information
- noisy images
- noise level
- gaussian processes
- denoising
- maximum likelihood
- expectation maximization
- model selection
- bayesian framework
- em algorithm
- natural images
- gaussian process
- sample size
- cross validation
- random sampling
- statistical tests
- closed form
- prior knowledge
- image processing
- support vector
- probabilistic model
- regularization parameter
- image reconstruction
- wavelet coefficients
- incomplete data
- markov random field
- confidence intervals
- posterior probability
- gaussian distribution
- parameter space
- data mining
- energy function
- learning algorithm