A closed-form solution to the graph total variation problem for continuous emotion profiling in noisy environment.
Shaoling JingXia MaoLijiang ChenMaria Colomba ComesArianna MencattiniGrazia RagusoFabien RingevalBjörn W. SchullerCorrado Di NataleEugenio MartinelliPublished in: Speech Commun. (2018)
Keyphrases
- total variation
- noisy environments
- image restoration
- denoising
- image denoising
- minimization problems
- convex optimization
- noise reduction
- blind deconvolution
- image deblurring
- regularization term
- speech recognition
- augmented lagrangian
- total variation minimization
- total variation regularization
- color image restoration
- image processing
- regularization methods
- fidelity term
- multiplicative noise removal
- speech signal
- closed form
- super resolution
- noisy images
- dual formulation
- natural images
- machine learning
- automatic speech recognition
- edge preserving
- weighted graph
- wavelet coefficients
- language model
- high quality