Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters
Abstract In this work, we proposed a method of extracting feature parameters for deep neural network prediction based on the vectorgraph storage format, which can be applied to the design of electromagnetic metamaterials with sandwich structures. Compared to current methods of manually extracting fe...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-03-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-32142-1 |
| _version_ | 1797853852145287168 |
|---|---|
| author | Jiale Gao Chunjie Feng Xingyi Wu Yanghui Wu Xiaobo Zhu Daying Sun Yutao Yue Wenhua Gu |
| author_facet | Jiale Gao Chunjie Feng Xingyi Wu Yanghui Wu Xiaobo Zhu Daying Sun Yutao Yue Wenhua Gu |
| author_sort | Jiale Gao |
| collection | DOAJ |
| description | Abstract In this work, we proposed a method of extracting feature parameters for deep neural network prediction based on the vectorgraph storage format, which can be applied to the design of electromagnetic metamaterials with sandwich structures. Compared to current methods of manually extracting feature parameters, this method can automatically and precisely extract the feature parameters of arbitrary two-dimensional surface patterns of the sandwich structure. The position and size of surface patterns can be freely defined, and the surface patterns can be easily scaled, rotated, translated, or transformed in other ways. Compared to the pixel graph feature extraction method, this method can adapt to very complex surface pattern design in a more efficient way. And the response band can be easily shifted by scaling the designed surface pattern. To illustrate and verify the method, a 7-layer deep neural network was built to design a metamaterial broadband polarization converter. Prototype samples were fabricated and tested to verify the accuracy of the prediction results. In general, the method is potentially applicable to the design of different kinds of sandwich-structure metamaterials, with different functions and in different frequency bands. |
| first_indexed | 2024-04-09T19:57:28Z |
| format | Article |
| id | doaj.art-67f687188099499eb303e29229418857 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-04-09T19:57:28Z |
| publishDate | 2023-03-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-67f687188099499eb303e292294188572023-04-03T05:23:04ZengNature PortfolioScientific Reports2045-23222023-03-0113111110.1038/s41598-023-32142-1Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization convertersJiale Gao0Chunjie Feng1Xingyi Wu2Yanghui Wu3Xiaobo Zhu4Daying Sun5Yutao Yue6Wenhua Gu7School of Microelectronics (School of Integrated Circuits), Nanjing University of Science and TechnologySchool of Microelectronics (School of Integrated Circuits), Nanjing University of Science and TechnologySchool of Microelectronics (School of Integrated Circuits), Nanjing University of Science and TechnologySchool of Microelectronics (School of Integrated Circuits), Nanjing University of Science and TechnologySchool of Microelectronics (School of Integrated Circuits), Nanjing University of Science and TechnologySchool of Microelectronics (School of Integrated Circuits), Nanjing University of Science and TechnologyInstitute of Deep Perception TechnologySchool of Microelectronics (School of Integrated Circuits), Nanjing University of Science and TechnologyAbstract In this work, we proposed a method of extracting feature parameters for deep neural network prediction based on the vectorgraph storage format, which can be applied to the design of electromagnetic metamaterials with sandwich structures. Compared to current methods of manually extracting feature parameters, this method can automatically and precisely extract the feature parameters of arbitrary two-dimensional surface patterns of the sandwich structure. The position and size of surface patterns can be freely defined, and the surface patterns can be easily scaled, rotated, translated, or transformed in other ways. Compared to the pixel graph feature extraction method, this method can adapt to very complex surface pattern design in a more efficient way. And the response band can be easily shifted by scaling the designed surface pattern. To illustrate and verify the method, a 7-layer deep neural network was built to design a metamaterial broadband polarization converter. Prototype samples were fabricated and tested to verify the accuracy of the prediction results. In general, the method is potentially applicable to the design of different kinds of sandwich-structure metamaterials, with different functions and in different frequency bands.https://doi.org/10.1038/s41598-023-32142-1 |
| spellingShingle | Jiale Gao Chunjie Feng Xingyi Wu Yanghui Wu Xiaobo Zhu Daying Sun Yutao Yue Wenhua Gu Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters Scientific Reports |
| title | Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters |
| title_full | Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters |
| title_fullStr | Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters |
| title_full_unstemmed | Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters |
| title_short | Deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters |
| title_sort | deep neural network training method based on vectorgraphs for designing of metamaterial broadband polarization converters |
| url | https://doi.org/10.1038/s41598-023-32142-1 |
| work_keys_str_mv | AT jialegao deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters AT chunjiefeng deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters AT xingyiwu deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters AT yanghuiwu deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters AT xiaobozhu deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters AT dayingsun deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters AT yutaoyue deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters AT wenhuagu deepneuralnetworktrainingmethodbasedonvectorgraphsfordesigningofmetamaterialbroadbandpolarizationconverters |