Multi-level Wavelet-CNN for Image Restoration

Liu, Pengju; Zhang, Hongzhi; Zhang, Kai; Li, Lin; Zuo, Wangmeng

doi:10.1109/cvprw.2018.00121

Cited by 623 publications

(359 citation statements)

References 53 publications

Supporting

Mentioning

357

Contrasting

Unclassified

Order By: Relevance

“…While in conventional U-Net, concatenation is adopted. Compared to our previous work [24], we have made several improvements such as: (i) Instead of directly decomposing input images by DWT, we first use conv blocks to extract features from input, which is empirically shown to be beneficial for image restoration. (ii) In the 3rd hierarchical level, we use more feature maps to enhance feature representation.…”

Section: A From Multi-level Wpt To Mwcnnmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Level Wavelet Convolutional Neural Networks

et al. 2019

Self Cite

View full text Add to dashboard Cite

In computer vision, convolutional networks (CNNs) often adopts pooling to enlarge receptive field which has the advantage of low computational complexity. However, pooling can cause information loss and thus is detrimental to further operations such as features extraction and analysis. Recently, dilated filter has been proposed to trade off between receptive field size and efficiency. But the accompanying gridding effect can cause a sparse sampling of input images with checkerboard patterns. To address this problem, in this paper, we propose a novel multi-level wavelet CNN (MWCNN) model to achieve better trade-off between receptive field size and computational efficiency. The core idea is to embed wavelet transform into CNN architecture to reduce the resolution of feature maps while at the same time, increasing receptive field. Specifically, MWCNN for image restoration is based on U-Net architecture, and inverse wavelet transform (IWT) is deployed to reconstruct the high resolution (HR) feature maps. The proposed MWCNN can also be viewed as an improvement of dilated filter and a generalization of average pooling, and can be applied to not only image restoration tasks, but also any CNNs requiring a pooling operation. The experimental results demonstrate effectiveness of the proposed MWCNN for tasks such as image denoising, single image super-resolution, JPEG image artifacts removal and object classification. The code and pre-trained models will be given at https://github.com/lpj-github-io/MWCNNv2. Index Terms-Convolutional networks, receptive field size, efficiency, multi-level wavelet.

show abstract

Section: A From Multi-level Wpt To Mwcnnmentioning

confidence: 99%

“…III-B on three representative image restoration tasks, respectively. Here, we also provide the results of our previous work and denote it as MWCNN(P) [24].…”

Section: B Quantitative and Qualitative Evaluation On Image Restoratmentioning

confidence: 99%

Multi-Level Wavelet Convolutional Neural Networks

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…it learns to isolate the noise H from the corrupted sample to later remove this noise instead of directly recovering the latent clean signal. Multi-level Wavelet Convolutionnal Neural Network (MWCNN) [3] is also CNN-based. Its novelty lies in the symmetrical use of wavelet and inverse wavelet transforms into the contracting and expanding parts of a U-Net [11] architecture.…”

Section: Related Work On Image Denoisersmentioning

confidence: 99%

“…If the noise is additive, H is a sum and y = x + η, where η is a realisation of H. There exist plenty of noise models to represent H among which examples of frequently used models are additive Gaussian, Poisson, salt and pepper or speckle noises. State of the art denoisers are constantly progressing in terms of noise elimination level [1,2,3]. However, most techniques are tailored for and evaluated on a given noise distribution, exploiting its probabilistic properties to distinguish it from the signal of interest.…”

Section: Introductionmentioning

confidence: 99%

Opendenoising: An Extensible Benchmark for Building Comparative Studies of Image Denoisers

Lemarchand

Montesuma

Pelcat

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Image denoising has recently taken a leap forward due to machine learning. However, image denoisers, both expert-based and learning-based, are mostly tested on well-behaved generated noises (usually Gaussian) rather than on real-life noises, making performance comparisons difficult in real-world conditions. This is especially true for learning-based denoisers which performance depends on training data. Hence, choosing which method to use for a specific denoising problem is difficult.This paper proposes a comparative study of existing denoisers, as well as an extensible open tool that makes it possible to reproduce and extend the study. MWCNN is shown to outperform other methods when trained for a real-world image interception noise, and additionally is the second least compute hungry of the tested methods. To evaluate the robustness of conclusions, three test sets are compared. A Kendall's Tau correlation of only 60% is obtained on methods ranking between noise types, demonstrating the need for a benchmarking tool.

show abstract

“…Texture classification using wavelet CNN [11], multi-scale face super-resolution [16], image super-resolution [13] and edge feature enhancing [10] are notable applications. A multi-level wavelet CNN model for image restoration has been introduced by Liu et al [21]. Cotter et al…”

Section: Related Workmentioning

confidence: 99%

Multi-path learnable wavelet neural network for image classification

Silva

Vithanage

Fernando³

et al. 2020

Twelfth International Conference on Machine Vision (ICMV 2019)

View full text Add to dashboard Cite

Despite the remarkable success of deep learning in pattern recognition, deep network models face the problem of training a large number of parameters. In this paper, we propose and evaluate a novel multi-path wavelet neural network architecture for image classification with far less number of trainable parameters. The model architecture consists of a multi-path layout with several levels of wavelet decompositions performed in parallel followed by fully connected layers. These decomposition operations comprise wavelet neurons with learnable parameters, which are updated during the training phase using the back-propagation algorithm. We evaluate the performance of the introduced network using common image datasets without data augmentation except for SVHN and compare the results with influential deep learning models. Our findings support the possibility of reducing the number of parameters significantly in deep neural networks without compromising its accuracy.

show abstract

Multi-level Wavelet-CNN for Image Restoration

Cited by 623 publications

References 53 publications

Multi-Level Wavelet Convolutional Neural Networks

Multi-Level Wavelet Convolutional Neural Networks

Opendenoising: An Extensible Benchmark for Building Comparative Studies of Image Denoisers

Multi-path learnable wavelet neural network for image classification

Contact Info

Product

Resources

About