Bayesian Restoration of Audio Degraded by Low-Frequency Pulses Modeled via Gaussian Process.
Hugo Tremonte de CarvalhoFlávio Rainho ÁvilaLuiz W. P. BiscainhoPublished in: IEEE J. Sel. Top. Signal Process. (2021)
Keyphrases
- low frequency
- gaussian process
- gaussian processes
- high frequency
- marginal likelihood
- frequency domain
- degraded images
- fully bayesian
- wavelet transform
- hyperparameters
- approximate inference
- gaussian process regression
- bayesian framework
- regression model
- model selection
- latent variables
- posterior distribution
- expectation propagation
- image restoration
- subband
- semi supervised
- maximum likelihood
- wavelet coefficients
- cross validation
- high resolution
- signal processing
- bayesian methods
- bayesian inference
- gaussian process models
- discrete wavelet transform
- high quality
- music information retrieval
- random sampling
- closed form
- probability distribution
- multiresolution
- feature vectors
- incomplete data
- machine learning
- feature extraction