Corpus ID: 235614202

GAN2GAN: Generative Noise Learning for Blind Denoising with Single Noisy Images

@inproceedings{Cha2021GAN2GANGN,
  title={GAN2GAN: Generative Noise Learning for Blind Denoising with Single Noisy Images},
  author={Sungmin Cha and Taeeon Park and Byeongjoon Kim and Jongduk Baek and Taesup Moon},
  booktitle={ICLR},
  year={2021}
}
We tackle a challenging blind image denoising problem, in which only single distinct noisy images are available for training a denoiser, and no information about noise is known, except for it being zero-mean, additive, and independent of the clean image. In such a setting, which often occurs in practice, it is not possible to train a denoiser with the standard discriminative training or with the recently developed Noise2Noise (N2N) training; the former requires the underlying clean image for… Expand
FBI-Denoiser: Fast Blind Image Denoiser for Poisson-Gaussian Noise
All experiments were conducted by eight-core Intel(R) Xeon(R) Silver 4110 CPU @ 2.10GHz, 128GB RAM and an NVIDIA Titan XP. Learning rages of 0.001 and 0.0001 were used for training FBI-Net andExpand
End-to-End Unsupervised Document Image Blind Denoising
TLDR
A unified end-toend unsupervised deep learning model is proposed, for the first time, that can effectively remove multiple types of noise, including salt & pepper noise, blurred and/or faded text, as well as watermarks from documents at various levels of intensity. Expand
Optimal Transport for Unsupervised Restoration Learning
TLDR
This work proposes a criterion for unsupervised restoration learning based on the optimal transport theory and shows remarkable superiority in harsh practical conditions with complex noise, e.g., raw depth images. Expand

References

SHOWING 1-10 OF 34 REFERENCES
Image Blind Denoising with Generative Adversarial Network Based Noise Modeling
TLDR
A novel two-step framework is proposed, in which a Generative Adversarial Network is trained to estimate the noise distribution over the input noisy images and to generate noise samples to train a deep Convolutional Neural Network for denoising. Expand
Adam: A Method for Stochastic Optimization
TLDR
This work introduces Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments, and provides a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework. Expand
Field of experts
Residual Dense Network for Image Restoration
TLDR
This work proposes residual dense block (RDB) to extract abundant local features via densely connected convolutional layers and proposes local feature fusion in RDB to adaptively learn more effective features from preceding and current local features and stabilize the training of wider network. Expand
A Poisson-Gaussian Denoising Dataset With Real Fluorescence Microscopy Images
  • Yide Zhang, Yinhao Zhu, +4 authors S. Howard
  • Computer Science, Engineering
  • 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2019
TLDR
This paper constructs a dataset - the Fluorescence Microscopy Denoising (FMD) dataset - that is dedicated to Poisson-Gaussian denoising and uses this dataset to benchmark 10 representative denoised algorithms and finds that deep learning methods have the best performance. Expand
High-Quality Self-Supervised Deep Image Denoising
TLDR
This work builds on a recent technique that removes the need for reference data by employing networks with a "blind spot" in the receptive field, and significantly improves two key aspects: image quality and training efficiency. Expand
Noise2Void - Learning Denoising From Single Noisy Images
TLDR
Noise2Void is introduced, a training scheme that allows us to train directly on the body of data to be denoised and can therefore be applied when other methods cannot, and compares favorably to training-free denoising methods. Expand
Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
TLDR
This paper investigates the construction of feed-forward denoising convolutional neural networks (DnCNNs) to embrace the progress in very deep architecture, learning algorithm, and regularization method into image Denoising, and uses residual learning and batch normalization to speed up the training process as well as boost theDenoising performance. Expand
Wasserstein Generative Adversarial Networks
TLDR
This work introduces a new algorithm named WGAN, an alternative to traditional GAN training that can improve the stability of learning, get rid of problems like mode collapse, and provide meaningful learning curves useful for debugging and hyperparameter searches. Expand
Noise2Noise: Learning Image Restoration without Clean Data
TLDR
It is shown that under certain common circumstances, it is possible to learn to restore signals without ever observing clean ones, at performance close or equal to training using clean exemplars. Expand
...
1
2
3
4
...