File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1016/j.neucom.2020.03.018
- WOS: WOS:000544724700011
- Find via
Supplementary
-
Citations:
- Web of Science: 0
- Appears in Collections:
Article: Deep plug-and-play prior for low-rank tensor completion
Title | Deep plug-and-play prior for low-rank tensor completion |
---|---|
Authors | |
Keywords | Tensor completion Tensor nuclear norm Denoising neural network Alternating direction method of multipliers Plug-and-play framework |
Issue Date | 2020 |
Publisher | ScienceDirect. The Journal's web site is located at http://www.elsevier.com/locate/neucom |
Citation | Neurocomputing, 2020, v. 400, p. 137-149 How to Cite? |
Abstract | Multi-dimensional images, such as color images and multi-spectral images (MSIs), are highly correlated and contain abundant spatial and spectral information. However, real-world multi-dimensional images are usually corrupted by missing entries. By integrating deterministic low-rankness prior to the data-driven deep prior, we suggest a novel regularized tensor completion model for multi-dimensional image completion. In the objective function, we adopt the newly emerged tensor nuclear norm (TNN) to characterize the global low-rankness prior of multi-dimensional images. We also formulate an implicit regularizer by plugging a denoising neural network (termed as deep denoiser), which is convinced to express the deep image prior learned from a large number of natural images. The resulting model can be solved by the alternating directional method of multipliers algorithm under the plug-and-play (PnP) framework. Experimental results on color images, videos, and MSIs demonstrate that the proposed method can recover both the global structure and fine details very well and achieve superior performance over competing methods in terms of quality metrics and visual effects. |
Persistent Identifier | http://hdl.handle.net/10722/304240 |
ISSN | 2023 Impact Factor: 5.5 2023 SCImago Journal Rankings: 1.815 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Zhao, X | - |
dc.contributor.author | Xu, W | - |
dc.contributor.author | Jiang, T | - |
dc.contributor.author | Wang, Y | - |
dc.contributor.author | Ng, KP | - |
dc.date.accessioned | 2021-09-23T08:57:13Z | - |
dc.date.available | 2021-09-23T08:57:13Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | Neurocomputing, 2020, v. 400, p. 137-149 | - |
dc.identifier.issn | 0925-2312 | - |
dc.identifier.uri | http://hdl.handle.net/10722/304240 | - |
dc.description.abstract | Multi-dimensional images, such as color images and multi-spectral images (MSIs), are highly correlated and contain abundant spatial and spectral information. However, real-world multi-dimensional images are usually corrupted by missing entries. By integrating deterministic low-rankness prior to the data-driven deep prior, we suggest a novel regularized tensor completion model for multi-dimensional image completion. In the objective function, we adopt the newly emerged tensor nuclear norm (TNN) to characterize the global low-rankness prior of multi-dimensional images. We also formulate an implicit regularizer by plugging a denoising neural network (termed as deep denoiser), which is convinced to express the deep image prior learned from a large number of natural images. The resulting model can be solved by the alternating directional method of multipliers algorithm under the plug-and-play (PnP) framework. Experimental results on color images, videos, and MSIs demonstrate that the proposed method can recover both the global structure and fine details very well and achieve superior performance over competing methods in terms of quality metrics and visual effects. | - |
dc.language | eng | - |
dc.publisher | ScienceDirect. The Journal's web site is located at http://www.elsevier.com/locate/neucom | - |
dc.relation.ispartof | Neurocomputing | - |
dc.subject | Tensor completion | - |
dc.subject | Tensor nuclear norm | - |
dc.subject | Denoising neural network | - |
dc.subject | Alternating direction method of multipliers | - |
dc.subject | Plug-and-play framework | - |
dc.title | Deep plug-and-play prior for low-rank tensor completion | - |
dc.type | Article | - |
dc.identifier.email | Ng, KP: michael.ng@hku.hk | - |
dc.identifier.authority | Ng, KP=rp02578 | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1016/j.neucom.2020.03.018 | - |
dc.identifier.hkuros | 325156 | - |
dc.identifier.volume | 400 | - |
dc.identifier.spage | 137 | - |
dc.identifier.epage | 149 | - |
dc.identifier.isi | WOS:000544724700011 | - |
dc.publisher.place | United States | - |