Universal Denoising Networks : A Novel CNN Architecture for Image Denoising

By Stamatios Lefkimmiatis. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2017

299 citations recorded by the source corpus.

We design a novel network architecture for learning discriminative image models that are employed to efficiently tackle the problem of grayscale and color image denoising. Based on the proposed architecture, we introduce two different variants. The first network involves convolutional layers as a core component, while the second one relies instead on non-local filtering layers and thus it is able to exploit the inherent non-local self-similarity property of natural images. As opposed to most of the existing deep network approaches, which require the training of a specific model for each considered noise level, the proposed models are able to handle a wide range of noise levels using a single set of learned parameters, while they are very robust when the noise degrading the latent image does not match the statistics of the noise used during training. The latter argument is supported by results that we report on publicly available images corrupted by unknown noise and which we compare against solutions obtained by competing methods. At the same time the introduced networks achieve excellent results under additive white Gaussian noise (AWGN), which are comparable to those of the current state-of-the-art network, while they depend on a more shallow architecture with the number of trained parameters being one order of magnitude smaller. These properties make the proposed networks ideal candidates to serve as sub-solvers on restoration methods that deal with general inverse imaging problems such as deblurring, demosaicking, superresolution, etc.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/635fe55f70feffca01fec7e07e4968b98dba672f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. Radu Timofte
2. Deyu Meng
3. Zongsheng Yue
4. Hongwei Yong
5. Qian Zhao
6. Mostafa El-Khamy
7. Jungwon Lee
8. Haoyu Ren

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Ming-Hsuan Yang
3. Radu Timofte
4. Lei Zhang
5. Martin Danelljan
6. Dacheng Tao
7. Fahad Shahbaz Khan
8. Wangmeng Zuo

In Media Technology

1. Radu Timofte
2. Lei Zhang
3. Kyung-ah Sohn
4. Namhyuk Ahn
5. Zongsheng Yue
6. Deyu Meng
7. Luc Van Gool
8. Martin Danelljan

Citations in the corpus

• Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Images (2021, background): “After that, numerous denoising networks are proposed to further improve the performance [20, 28, 35, 16, 24, 12, 10].”
• NTIRE 2020 Challenge on Real-World Image Super-Resolution: Methods and Results (2020, methodology): “The trainable projection layer [42] inside Decoder computes the proximal map with the estimated noise standard deviation and handles the data fidelity and prior terms.”
• Variational Denoising Network: Toward Blind Noise Modeling and Removal (2019, methodology): “From Table 1, it can be easily observed that: 1) The VDN outperforms other competing methods in all cases, indicating that VDN is able to handle such complicated non-i.i.d. noise; 2) VDN surpasses FFDNet about 0.25dB averagely even though FFDNet depends on the true noise level information instead of”
• Noise2Void - Learning Denoising From Single Noisy Images (2018, background): “In recent years, convolutional neural networks (CNNs) have been trained in various ways to predict pixel values from surrounding image patches, i. tive field of that pixel [24, 11, 26, 6, 23, 25, 18, 14].”