R3L: Connecting Deep Reinforcement Learning To Recurrent Neural Networks For Image Denoising Via Residual Recovery.
Rongkai ZhangJiang ZhuZhiyuan ZhaJustin DauwelsBihan WenPublished in: ICIP (2021)
Keyphrases
- recurrent neural networks
- image denoising
- reinforcement learning
- denoising
- natural images
- noisy images
- total variation
- neural network
- feed forward
- image processing
- partial differential equations
- recurrent networks
- echo state networks
- wavelet coefficients
- wavelet domain
- bilateral filtering
- nonlocal means
- wavelet domain hidden markov tree
- image deblurring
- denoising algorithm
- artificial neural networks
- undecimated wavelet transform
- denoising methods
- additive gaussian noise
- wavelet denoising
- machine learning
- image inpainting
- input image
- multiscale
- learning algorithm
- nonlinear dynamic systems