Transfer Learning From Synthetic to Real-Noise Denoising With Adaptive Instance Normalization

  title={Transfer Learning From Synthetic to Real-Noise Denoising With Adaptive Instance Normalization},
  author={Yoonsik Kim and Jae Woong Soh and Gu Yong Park and Nam Ik Cho},
  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  • Yoonsik KimJae Woong Soh N. Cho
  • Published 26 February 2020
  • Computer Science
  • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Real-noise denoising is a challenging task because the statistics of real-noise do not follow the normal distribution, and they are also spatially and temporally changing. In order to cope with various and complex real-noise, we propose a well-generalized denoising architecture and a transfer learning scheme. Specifically, we adopt an adaptive instance normalization to build a denoiser, which can regularize the feature map and prevent the network from overfitting to the training set. We also… 

C2N: Practical Generative Noise Modeling for Real-World Denoising

A Clean-to-Noisy image generation framework, namely C2N, to imitate complex real-world noise without using any paired examples, and conventional denoising CNNs can be trained to outperform existing unsupervised methods on challenging real- world benchmarks by a large margin.

Generative Adaptive Convolutions for Real-World Noisy Image Denoising

This work proposes a novel flexible and adaptive denoising network, coined as FADNet, equipped with a plane dynamic filter module, which generates weight filters with flexibility that can adapt to the specific input and thereby impedes the FAD net from overfitting to the training data.

Adaptive noise imitation for image denoising

A new Adaptive noise imitation (ADANI) algorithm is developed that can synthesize noisy data from naturally noisy images and is competitive to other networks trained with external paired data.

Invertible Denoising Network: A Light Solution for Real Noise Removal

The proposed invertible denoising network, InvDN, is better than all the existing competitive models, achieving a new state-of-the-art result for the SIDD dataset while enjoying less run time.

Learning to Generate Realistic Noisy Images via Pixel-level Noise-aware Adversarial Training

Qualitative validation shows that noise generated by PNGAN is highly similar to real noise in terms of intensity and distribution, which clearly suggests the high similarity between PNGAN generating and real noisy images.

Disentangling Noise from Images: A Flow-Based Image Denoising Neural Network

An invertible denoising network, FDN, without any assumptions on either clean or noise distributions, as well as a distribution disentanglement method, which learns the distribution of noisy images, which is different from the previous CNN based discriminative mapping.

Controllable Confidence-Based Image Denoising

This project presents a framework that is capable of controllable, confidence-based noise removal, based on the fusion between two different denoised images, both derived from the same noisy input.

Self-supervised training for blind multi-frame video denoising

This work uses the proposed strategy to denoise a video contaminated with an unknown noise type, by fine-tuning a pre-trained denoising network on the noisy video, and shows how this can be expoited to perform jointDenoising and noise level estimation for heteroscedastic noise.

Variational Deep Image Denoising

A novel Bayesian framework based on the variational approximation of objective functions is proposed that provides remarkable performance on additive white Gaussian noise and real-noise denoising while requiring fewer parameters than recent state-of-the-art denoisers.

Unsupervised Image Denoising with Frequency Domain Knowledge

Results using natural and synthetic datasets indicate that the unsupervised learning method augmented with frequency information achieves state-of-the-art denoising performance, suggesting that frequency domain information could be a viable factor in improving the overall performance of un supervised learning-based methods.



Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising

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.

Toward Convolutional Blind Denoising of Real Photographs

A convolutional blind denoising network (CBDNet) with more realistic noise model and real-world noisy-clean image pairs and a noise estimation subnetwork with asymmetric learning to suppress under-estimation of noise level is embedded into CBDNet.

GRDN:Grouped Residual Dense Network for Real Image Denoising and GAN-Based Real-World Noise Modeling

A grouped residual dense network (GRDN) is proposed, which is an extended and generalized architecture of the state-of-the-art residual densenetwork (RDN) and a new generative adversarial network-based real-world noise modeling method is developed.

FFDNet: Toward a Fast and Flexible Solution for CNN-Based Image Denoising

The proposed FFDNet works on downsampled sub-images, achieving a good trade-off between inference speed and denoising performance, and enjoys several desirable properties, including the ability to handle a wide range of noise levels effectively with a single network.

Image Blind Denoising with Generative Adversarial Network Based Noise Modeling

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.

Universal Denoising Networks : A Novel CNN Architecture for Image Denoising

A novel network architecture for learning discriminative image models that are employed to efficiently tackle the problem of grayscale and color image denoising is designed and two different variants are introduced, which achieve excellent results under additive white Gaussian noise.

Real-world Noisy Image Denoising: A New Benchmark

A new benchmark dataset which contains comprehensive real-world noisy images of different natural scenes and demonstrates that the recently proposed methods designed specifically for realistic noise removal based on sparse or low rank theories achieve better denoising performance and are more robust than other competing methods.

Noise Flow: Noise Modeling With Conditional Normalizing Flows

Noise Flow is introduced, a powerful and accurate noise model based on recent normalizing flow architectures that represents the first serious attempt to go beyond simple parametric models to one that leverages the power of deep learning and data-driven noise distributions.

Adaptively Tuning a Convolutional Neural Network by Gate Process for Image Denoising

A newDenoising scheme is proposed that controls the feature maps of a single denoising network according to the noise level at the test phase, without changing the network parameters.

RENOIR - A dataset for real low-light image noise reduction