Improved wavelet prediction superresolution reconstruction based on U‐Net
Abstract Deep learning can be used to achieve single‐image superresolution (SR) reconstruction. To address problems encountered during this process, such as the number of network parameters, high training requirements on equipment performance, and inability to downsample certain SR images accurately...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2023-10-01
|
| Series: | IET Image Processing |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/ipr2.12878 |
| _version_ | 1797666034870648832 |
|---|---|
| author | Jianfang Cao Zeyu Chen Hongyan Cui Xiaofei Ji Xianhui Wang Yunchuan Liang Yun Tian |
| author_facet | Jianfang Cao Zeyu Chen Hongyan Cui Xiaofei Ji Xianhui Wang Yunchuan Liang Yun Tian |
| author_sort | Jianfang Cao |
| collection | DOAJ |
| description | Abstract Deep learning can be used to achieve single‐image superresolution (SR) reconstruction. To address problems encountered during this process, such as the number of network parameters, high training requirements on equipment performance, and inability to downsample certain SR images accurately, an image SR reconstruction algorithm based on deep residual network optimization is proposed. The model introduces wavelet transforms based on the original U‐Net, where the U‐Net is trained to obtain SR wavelet feature images at multiple scales simultaneously. This approach reduces the mapping space for the network to learn low‐ to high‐resolution image mapping, which in turn reduces the training difficulty of the model. In terms of network details, the inverse wavelet transform is used in image upsampling to enhance the sparsity of the reconstruction layer in the original network. The network structure of the U‐Net upsampling is adjusted slightly to enable the network to distinguish wavelet images from feature images, thereby improving the richness of the features extracted by the model. The experimental results show that the peak signal‐to‐noise ratio (PSNR) of the fourfold SR model is 32.35 and 28.68 on the Set5 and Set14 validation sets, respectively. Compared with networks that use wavelet prediction mechanisms, such as the deep wavelet prediction SR (DWSR) and deep wavelet prediction‐based residual SR (DWRSR) models, the PSNR for all the tested public datasets is improved by 0.5. The method yields superior results in terms of both visual effect and PSNR, demonstrating the feasibility of the wavelet prediction mechanism in SR reconstruction and thus offering application value and research significance. |
| first_indexed | 2024-03-11T19:52:24Z |
| format | Article |
| id | doaj.art-7dc56c7747064b258a80ec27a5b180ce |
| institution | Directory Open Access Journal |
| issn | 1751-9659 1751-9667 |
| language | English |
| last_indexed | 2024-03-11T19:52:24Z |
| publishDate | 2023-10-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Image Processing |
| spelling | doaj.art-7dc56c7747064b258a80ec27a5b180ce2023-10-05T03:50:15ZengWileyIET Image Processing1751-96591751-96672023-10-0117123464347610.1049/ipr2.12878Improved wavelet prediction superresolution reconstruction based on U‐NetJianfang Cao0Zeyu Chen1Hongyan Cui2Xiaofei Ji3Xianhui Wang4Yunchuan Liang5Yun Tian6School of Computer Science and Technology Taiyuan University of Science and Technology TaiyuanChinaSchool of Computer Science and Technology Taiyuan University of Science and Technology TaiyuanChinaSchool of Computer Science and Technology Taiyuan University of Science and Technology TaiyuanChinaDepartment of Computer Science and Technology Xinzhou Normal University XinzhouChinaSchool of Computer Science and Technology Taiyuan University of Science and Technology TaiyuanChinaDepartment of Computer Science and Technology Xinzhou Normal University XinzhouChinaDepartment of Computer Science and Technology Xinzhou Normal University XinzhouChinaAbstract Deep learning can be used to achieve single‐image superresolution (SR) reconstruction. To address problems encountered during this process, such as the number of network parameters, high training requirements on equipment performance, and inability to downsample certain SR images accurately, an image SR reconstruction algorithm based on deep residual network optimization is proposed. The model introduces wavelet transforms based on the original U‐Net, where the U‐Net is trained to obtain SR wavelet feature images at multiple scales simultaneously. This approach reduces the mapping space for the network to learn low‐ to high‐resolution image mapping, which in turn reduces the training difficulty of the model. In terms of network details, the inverse wavelet transform is used in image upsampling to enhance the sparsity of the reconstruction layer in the original network. The network structure of the U‐Net upsampling is adjusted slightly to enable the network to distinguish wavelet images from feature images, thereby improving the richness of the features extracted by the model. The experimental results show that the peak signal‐to‐noise ratio (PSNR) of the fourfold SR model is 32.35 and 28.68 on the Set5 and Set14 validation sets, respectively. Compared with networks that use wavelet prediction mechanisms, such as the deep wavelet prediction SR (DWSR) and deep wavelet prediction‐based residual SR (DWRSR) models, the PSNR for all the tested public datasets is improved by 0.5. The method yields superior results in terms of both visual effect and PSNR, demonstrating the feasibility of the wavelet prediction mechanism in SR reconstruction and thus offering application value and research significance.https://doi.org/10.1049/ipr2.12878channel attentionconvolutional neural networkdeep learningsingle‐image superresolutionwavelet transform |
| spellingShingle | Jianfang Cao Zeyu Chen Hongyan Cui Xiaofei Ji Xianhui Wang Yunchuan Liang Yun Tian Improved wavelet prediction superresolution reconstruction based on U‐Net IET Image Processing channel attention convolutional neural network deep learning single‐image superresolution wavelet transform |
| title | Improved wavelet prediction superresolution reconstruction based on U‐Net |
| title_full | Improved wavelet prediction superresolution reconstruction based on U‐Net |
| title_fullStr | Improved wavelet prediction superresolution reconstruction based on U‐Net |
| title_full_unstemmed | Improved wavelet prediction superresolution reconstruction based on U‐Net |
| title_short | Improved wavelet prediction superresolution reconstruction based on U‐Net |
| title_sort | improved wavelet prediction superresolution reconstruction based on u net |
| topic | channel attention convolutional neural network deep learning single‐image superresolution wavelet transform |
| url | https://doi.org/10.1049/ipr2.12878 |
| work_keys_str_mv | AT jianfangcao improvedwaveletpredictionsuperresolutionreconstructionbasedonunet AT zeyuchen improvedwaveletpredictionsuperresolutionreconstructionbasedonunet AT hongyancui improvedwaveletpredictionsuperresolutionreconstructionbasedonunet AT xiaofeiji improvedwaveletpredictionsuperresolutionreconstructionbasedonunet AT xianhuiwang improvedwaveletpredictionsuperresolutionreconstructionbasedonunet AT yunchuanliang improvedwaveletpredictionsuperresolutionreconstructionbasedonunet AT yuntian improvedwaveletpredictionsuperresolutionreconstructionbasedonunet |