Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network
Deep neural networks (DNNs) have been widely adopted in single image super-resolution (SISR) recently with great success. As a network goes deeper, intermediate features become hierarchical. However, most SISR methods based on DNNs do not make full use of the hierarchical features. The features cann...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-01-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1424-8220/19/2/316 |
| _version_ | 1817988958311677952 |
|---|---|
| author | Wang Xu Renwen Chen Bin Huang Xiang Zhang Chuan Liu |
| author_facet | Wang Xu Renwen Chen Bin Huang Xiang Zhang Chuan Liu |
| author_sort | Wang Xu |
| collection | DOAJ |
| description | Deep neural networks (DNNs) have been widely adopted in single image super-resolution (SISR) recently with great success. As a network goes deeper, intermediate features become hierarchical. However, most SISR methods based on DNNs do not make full use of the hierarchical features. The features cannot be read directly by the subsequent layers, therefore, the previous hierarchical information has little influence on the subsequent layer output, and the performance is relatively poor. To address this issue, a novel global dense feature fusion convolutional network (DFFNet) is proposed, which can take full advantage of global intermediate features. Especially, a feature fusion block (FFblock) is introduced as the basic module. Each block can directly read raw global features from previous ones and then learns the feature spatial correlation and channel correlation between features in a holistic way, leading to a continuous global information memory mechanism. Experiments on the benchmark tests show that the proposed method DFFNet achieves favorable performance against the state-of-art methods. |
| first_indexed | 2024-04-14T00:41:15Z |
| format | Article |
| id | doaj.art-20477941c6f84a048b0ccf662ce1ad81 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-14T00:41:15Z |
| publishDate | 2019-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-20477941c6f84a048b0ccf662ce1ad812022-12-22T02:22:11ZengMDPI AGSensors1424-82202019-01-0119231610.3390/s19020316s19020316Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional NetworkWang Xu0Renwen Chen1Bin Huang2Xiang Zhang3Chuan Liu4State Key Laboratory of Mechanics and Control of Mechanical Structures, College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, ChinaState Key Laboratory of Mechanics and Control of Mechanical Structures, College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, ChinaState Key Laboratory of Mechanics and Control of Mechanical Structures, College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, ChinaState Key Laboratory of Mechanics and Control of Mechanical Structures, College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, ChinaState Key Laboratory of Mechanics and Control of Mechanical Structures, College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, ChinaDeep neural networks (DNNs) have been widely adopted in single image super-resolution (SISR) recently with great success. As a network goes deeper, intermediate features become hierarchical. However, most SISR methods based on DNNs do not make full use of the hierarchical features. The features cannot be read directly by the subsequent layers, therefore, the previous hierarchical information has little influence on the subsequent layer output, and the performance is relatively poor. To address this issue, a novel global dense feature fusion convolutional network (DFFNet) is proposed, which can take full advantage of global intermediate features. Especially, a feature fusion block (FFblock) is introduced as the basic module. Each block can directly read raw global features from previous ones and then learns the feature spatial correlation and channel correlation between features in a holistic way, leading to a continuous global information memory mechanism. Experiments on the benchmark tests show that the proposed method DFFNet achieves favorable performance against the state-of-art methods.http://www.mdpi.com/1424-8220/19/2/316dense feature fusionconvolutional neural networkimage super-resolution |
| spellingShingle | Wang Xu Renwen Chen Bin Huang Xiang Zhang Chuan Liu Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network Sensors dense feature fusion convolutional neural network image super-resolution |
| title | Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network |
| title_full | Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network |
| title_fullStr | Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network |
| title_full_unstemmed | Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network |
| title_short | Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network |
| title_sort | single image super resolution based on global dense feature fusion convolutional network |
| topic | dense feature fusion convolutional neural network image super-resolution |
| url | http://www.mdpi.com/1424-8220/19/2/316 |
| work_keys_str_mv | AT wangxu singleimagesuperresolutionbasedonglobaldensefeaturefusionconvolutionalnetwork AT renwenchen singleimagesuperresolutionbasedonglobaldensefeaturefusionconvolutionalnetwork AT binhuang singleimagesuperresolutionbasedonglobaldensefeaturefusionconvolutionalnetwork AT xiangzhang singleimagesuperresolutionbasedonglobaldensefeaturefusionconvolutionalnetwork AT chuanliu singleimagesuperresolutionbasedonglobaldensefeaturefusionconvolutionalnetwork |