Provable ICA with Unknown Gaussian Noise, and Implications for Gaussian Mixtures and Autoencoders.
Sanjeev AroraRong GeAnkur MoitraSushant SachdevaPublished in: Algorithmica (2015)
Keyphrases
- gaussian noise
- gaussian mixture
- denoising
- noisy images
- image denoising
- closed form
- noise removal
- em algorithm
- gaussian mixture model
- expectation maximization
- image processing
- covariance matrix
- probability density function
- image restoration
- natural images
- density estimation
- impulse noise
- mixture model
- signal to noise ratio
- noise level
- salt and pepper noise
- face recognition
- feature extraction
- signal processing
- image segmentation
- feature selection
- statistical analysis
- maximum likelihood
- principal component analysis
- machine learning