Abstract
Despite the significant advances in convolutional neural network (CNN) based image denoising, the existing methods still cannot consistently outperform non-local self-similarity (NSS) based methods, especially on images with many repetitive structures. Although several studies have been given to incorporate NSS priors with CNN-based denoising,their improvement is generally insignificant when compared with the state-of-the-art CNN-based denoisers. In this paper, we suggest to combine CNN and NSS based methods for improved image denoising, resulting in an NSS-UNet architecture. Motivated by gradient descent inference of TNRD, both the current estimate and noisy observation are considered as the inputs to the CNN. To take the NSS prior into account, the result by NSS (e.g., BM3D or WNNM), is adopted as the initial estimate. And a modified UNet is presented for exploiting the multi-scale information. We evaluate the proposed method on three common testing datasets. The results clearly show that NSS-UNet outperforms the existing CNN and NSS based methods in terms of both PSNR index and visual quality.
Highlights
Image denoising is a fundamental and classical topic in image processing and low level vision
A wide range of models have been suggested for image denoising, to name a few, variational models [1]–[3], non-local self-similarity (NSS) based methods [4]–[6], Markov random fields [7]–[9], sparse representation methods [4]–[6], [10]–[12] and discriminative learning based methods [13]–[15]
By analyzing the connection with trainable non-linear reaction diffusion (TNRD), we argue that both the initial estimate xt and the noisy observation y should be taken as the inputs to the convolutional neural network (CNN) model
Summary
Image denoising is a fundamental and classical topic in image processing and low level vision. Extensive experiments show that our NSS-UNet outperforms the state-of-the-art model-based denoising methods, and even achieve denoising results comparable with current best performing NSS and CNN based methods in terms of quantitative and qualitative evaluation. CNN-based methods (e.g., DnCNN [21]) have achieved state-of-the-art denoising performance, they cannot consistently outperform the NSS-based methods, especially on images with many repetitive structures.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
