Learning Deep CNN Denoiser Prior for Image Restoration

Zhang, Kai; Zuo, Wangmeng; Gu, Shuhang; Zhang, Lei

doi:10.1109/cvpr.2017.300

Cited by 1,569 publications

(1,263 citation statements)

References 61 publications

Supporting

Mentioning

1,258

Contrasting

Unclassified

Order By: Relevance

“…The augmentation for test images further boosted the performance. We report the CNN [29] with the same augmentation for a fair comparison. It demonstrates that the pixel-wise combination of traditional image filtering methods can perform better than classical sophisticated denoising 0 2,000,000 4,000,000 6,000,000 8,000,000 10,000,000 12,000,000 methods (BM3D [23] and WNNM [58]) and comparative to state-of-the-art NN-based methods (TNRD [59], MLP [28], and CNN [29]).…”

Section: Applications and Resultsmentioning

confidence: 99%

PixelRL: Fully Convolutional Network With Reinforcement Learning for Image Processing

Furuta

Inoue

Yamasaki

2020

IEEE Trans. Multimedia

View full text Add to dashboard Cite

This paper tackles a new problem setting: reinforcement learning with pixel-wise rewards (pixelRL) for image processing. After the introduction of the deep Q-network, deep RL has been achieving great success. However, the applications of deep reinforcement learning (RL) for image processing are still limited. Therefore, we extend deep RL to pixelRL for various image processing applications. In pixelRL, each pixel has an agent, and the agent changes the pixel value by taking an action. We also propose an effective learning method for pixelRL that significantly improves the performance by considering not only the future states of the own pixel but also those of the neighbor pixels. The proposed method can be applied to some image processing tasks that require pixel-wise manipulations, where deep RL has never been applied. Besides, it is possible to visualize what kind of operation is employed for each pixel at each iteration, which would help us understand why and how such an operation is chosen. We also believe that our technology can enhance the explainability and interpretability of the deep neural networks. In addition, because the operations executed at each pixels are visualized, we can change or modify the operations if necessary.We apply the proposed method to a variety of image processing tasks: image denoising, image restoration, local color enhancement, and saliency-driven image editing. Our experimental results demonstrate that the proposed method achieves comparable or better performance, compared with the state-ofthe-art methods based on supervised learning. The source code is available on https://github.com/rfuruta/pixelRL.

show abstract

Section: Applications and Resultsmentioning

confidence: 99%

PixelRL: Fully Convolutional Network With Reinforcement Learning for Image Processing

Furuta

Inoue

Yamasaki

2020

IEEE Trans. Multimedia

View full text Add to dashboard Cite

show abstract

“…In Dietrich et al [27], the convolutional neural network model was applied to image deburring. CNN have also achieved good results in image denoising [28,29].…”

Section: Introductionmentioning

confidence: 88%

A Deconvolution Technology of Microwave Radiometer Data Using Convolutional Neural Networks

Zhang

Shi

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

Abstract:Microwave radiometer data is affected by many factors during the imaging process, including the antenna pattern, system noise, and the curvature of the Earth. Existing deconvolution methods such as Wiener filtering handle this degradation problem in the Fourier domain. However, under complex degradation conditions, the Wiener filtering results are not accurate. In this paper, a convolutional neural network (CNN) model is proposed to solve the degradation problem. The deconvolution procedure is defined as a regression problem in the spatial domain that can be solved with deep learning. For the real inverse process of microwave radiometer data, the CNN model has a more powerful reconstruction ability than Wiener filtering due to the multi-layer structure of the CNN, which enables the multiple feature transform of the data. Additionally, the complex degradation factor during the imaging process of a microwave radiometer can be solved with general framework-based learning. Experimental results demonstrated that the CNN model gains about 5 dB at the peak signal-to-noise ratio compared to the Wiener filtering deconvolution method, and can better distinguish the measured data.

show abstract

“…Amortized or compiled inference attempt to directly learn an approximationF −1 (y) 1 for the problem y F (x), often by training a network with a very large number of known (x, y) pairs. In the applications highlighted in [26], including superresolution imaging [22], motion deblurring [49], and denoising [50], the output ofF −1 is a dense image. Other recent work [21] has attempted to learn an inverse model to classify the hand-written MNIST dataset from a camera system with minimal optics.…”

Section: Function Approximation Methods For Inverse Problemsmentioning

confidence: 99%

DeepLoco: Fast 3D Localization Microscopy Using Neural Networks

Boyd¹,

Jonas²,

Babcock³

et al. 2018

Preprint

View full text Add to dashboard Cite

Single-molecule localization super-resolution microscopy (SMLM) techniques like STORM and PALM have transformed cellular microscopy by substantially increasing spatial resolution. In this paper we introduce a new algorithm for a critical part of the SMLM process: estimating the number and locations of the fluorophores in a single frame. Our algorithm can analyze a 20000-frame experimental 3D SMLM dataset in about one second -substantially faster than real-time and existing algorithms. Our approach is straightforward but very different from existing algorithms: we train a neural network to minimize the Bayes' risk under a generative model for single SMLM frames. The neural network maps a frame directly to a collection of fluorophore locations, which we compare to the ground truth using a novel loss function. While training the neural network takes several hours, it only has to be done once for a given experimental setup. After training, localizing fluorophores in new images is extremely fast -orders of magnitude faster than existing algorithms. Faster recovery opens the door to real-time calibration and accelerated acquisition, and future work could tackle more complicated optical systems and more realistic simulators.

show abstract

Learning Deep CNN Denoiser Prior for Image Restoration

Cited by 1,569 publications

References 61 publications

PixelRL: Fully Convolutional Network With Reinforcement Learning for Image Processing

PixelRL: Fully Convolutional Network With Reinforcement Learning for Image Processing

A Deconvolution Technology of Microwave Radiometer Data Using Convolutional Neural Networks

DeepLoco: Fast 3D Localization Microscopy Using Neural Networks

Contact Info

Product

Resources

About