Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-Resolution
Single image super-resolution is known to be an ill-posed problem, which has been studied for decades. With the developments of deep convolutional neural networks, the CNN-based single image super-resolution methods have greatly improved the quality of the generated high-resolution images. However,...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8712146/ |
| _version_ | 1819133181745430528 |
|---|---|
| author | Chuangchuang Liu Xianfang Sun Changyou Chen Paul L. Rosin Yitong Yan Longcun Jin Xinyi Peng |
| author_facet | Chuangchuang Liu Xianfang Sun Changyou Chen Paul L. Rosin Yitong Yan Longcun Jin Xinyi Peng |
| author_sort | Chuangchuang Liu |
| collection | DOAJ |
| description | Single image super-resolution is known to be an ill-posed problem, which has been studied for decades. With the developments of deep convolutional neural networks, the CNN-based single image super-resolution methods have greatly improved the quality of the generated high-resolution images. However, it is difficult for image super-resolution to make full use of the relationship between pixels in low-resolution images. To address this issue, we propose a novel multi-scale residual hierarchical dense network, which tries to find the dependencies in multi-level and multi-scale features. Especially, we apply the atrous spatial pyramid pooling, which concatenates multiple atrous convolutions with different dilation rates, and design a residual hierarchical dense structure for single image super-resolution. The atrous-spatial-pyramid-pooling module is used for learning the relationship of features at multiple scales while the residual hierarchical dense structure, which consists of several hierarchical dense blocks with skip connections, aims to adaptively detect key information from multi-level features. Meanwhile, dense features from different groups are connected in a dense approach by hierarchical dense blocks, which can adequately extract local multi-level features. The extensive experiments on benchmark datasets illustrate the superiority of our proposed method compared with the state-of-the-art methods. The super-resolution results on benchmark datasets of our method can be downloaded from https://github.com/Rainyfish/MS-RHDN, and the source code will be released upon acceptance of the paper. |
| first_indexed | 2024-12-22T09:43:13Z |
| format | Article |
| id | doaj.art-347bcd6e4caa424690ca68f03f81c800 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-22T09:43:13Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-347bcd6e4caa424690ca68f03f81c8002022-12-21T18:30:36ZengIEEEIEEE Access2169-35362019-01-017605726058310.1109/ACCESS.2019.29159438712146Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-ResolutionChuangchuang Liu0Xianfang Sun1https://orcid.org/0000-0002-6114-0766Changyou Chen2Paul L. Rosin3Yitong Yan4Longcun Jin5https://orcid.org/0000-0002-1300-345XXinyi Peng6School of Software Engineering, South China University of Technology, Guangzhou, ChinaSchool of Computer Science and Informatics, Cardiff University, Cardiff, U.K.Department of Computer Science and Engineering, University at Buffalo, The State University of New York, Buffalo, NY, USASchool of Computer Science and Informatics, Cardiff University, Cardiff, U.K.School of Software Engineering, South China University of Technology, Guangzhou, ChinaSchool of Software Engineering, South China University of Technology, Guangzhou, ChinaSchool of Software Engineering, South China University of Technology, Guangzhou, ChinaSingle image super-resolution is known to be an ill-posed problem, which has been studied for decades. With the developments of deep convolutional neural networks, the CNN-based single image super-resolution methods have greatly improved the quality of the generated high-resolution images. However, it is difficult for image super-resolution to make full use of the relationship between pixels in low-resolution images. To address this issue, we propose a novel multi-scale residual hierarchical dense network, which tries to find the dependencies in multi-level and multi-scale features. Especially, we apply the atrous spatial pyramid pooling, which concatenates multiple atrous convolutions with different dilation rates, and design a residual hierarchical dense structure for single image super-resolution. The atrous-spatial-pyramid-pooling module is used for learning the relationship of features at multiple scales while the residual hierarchical dense structure, which consists of several hierarchical dense blocks with skip connections, aims to adaptively detect key information from multi-level features. Meanwhile, dense features from different groups are connected in a dense approach by hierarchical dense blocks, which can adequately extract local multi-level features. The extensive experiments on benchmark datasets illustrate the superiority of our proposed method compared with the state-of-the-art methods. The super-resolution results on benchmark datasets of our method can be downloaded from https://github.com/Rainyfish/MS-RHDN, and the source code will be released upon acceptance of the paper.https://ieeexplore.ieee.org/document/8712146/Convolutional neural networksdeep learningmulti-scale residual hierarchical denseimage super-resolution |
| spellingShingle | Chuangchuang Liu Xianfang Sun Changyou Chen Paul L. Rosin Yitong Yan Longcun Jin Xinyi Peng Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-Resolution IEEE Access Convolutional neural networks deep learning multi-scale residual hierarchical dense image super-resolution |
| title | Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-Resolution |
| title_full | Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-Resolution |
| title_fullStr | Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-Resolution |
| title_full_unstemmed | Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-Resolution |
| title_short | Multi-Scale Residual Hierarchical Dense Networks for Single Image Super-Resolution |
| title_sort | multi scale residual hierarchical dense networks for single image super resolution |
| topic | Convolutional neural networks deep learning multi-scale residual hierarchical dense image super-resolution |
| url | https://ieeexplore.ieee.org/document/8712146/ |
| work_keys_str_mv | AT chuangchuangliu multiscaleresidualhierarchicaldensenetworksforsingleimagesuperresolution AT xianfangsun multiscaleresidualhierarchicaldensenetworksforsingleimagesuperresolution AT changyouchen multiscaleresidualhierarchicaldensenetworksforsingleimagesuperresolution AT paullrosin multiscaleresidualhierarchicaldensenetworksforsingleimagesuperresolution AT yitongyan multiscaleresidualhierarchicaldensenetworksforsingleimagesuperresolution AT longcunjin multiscaleresidualhierarchicaldensenetworksforsingleimagesuperresolution AT xinyipeng multiscaleresidualhierarchicaldensenetworksforsingleimagesuperresolution |