R3L: Connecting Deep Reinforcement Learning to Recurrent Neural Networks for Image Denoising via Residual Recovery.
Rongkai ZhangJiang ZhuZhiyuan ZhaJustin DauwelsBihan WenPublished in: CoRR (2021)
Keyphrases
- image denoising
- recurrent neural networks
- reinforcement learning
- denoising
- noisy images
- natural images
- partial differential equations
- neural network
- feed forward
- total variation
- recurrent networks
- image processing
- echo state networks
- artificial neural networks
- wavelet coefficients
- image deblurring
- nonlocal means
- wavelet domain
- wavelet domain hidden markov tree
- additive gaussian noise
- bilateral filtering
- denoising methods
- machine learning
- augmented lagrangian
- learning algorithm
- total variation regularization
- pairwise
- computer vision