Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral Interference
Deep neural networks have enabled the reconstruction of optical soliton molecules with more complex structures using the real-time spectral interferences obtained by photonic time-stretch dispersive Fourier transformation (TS-DFT) technology. In this paper, we propose to use three kinds of deep conv...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-02-01
|
| Series: | Photonics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2304-6732/8/2/51 |
| _version_ | 1797396571819606016 |
|---|---|
| author | Caiyun Li Jiangyong He Yange Liu Yang Yue Luhe Zhang Longfei Zhu Mengjie Zhou Congcong Liu Kaiyan Zhu Zhi Wang |
| author_facet | Caiyun Li Jiangyong He Yange Liu Yang Yue Luhe Zhang Longfei Zhu Mengjie Zhou Congcong Liu Kaiyan Zhu Zhi Wang |
| author_sort | Caiyun Li |
| collection | DOAJ |
| description | Deep neural networks have enabled the reconstruction of optical soliton molecules with more complex structures using the real-time spectral interferences obtained by photonic time-stretch dispersive Fourier transformation (TS-DFT) technology. In this paper, we propose to use three kinds of deep convolution networks (DCNs), including VGG, ResNets, and DenseNets, for revealing internal dynamics evolution of soliton molecules based on the real-time spectral interferences. When analyzing soliton molecules with equidistant composite structures, all three models are effective. The DenseNets with layers of 48 perform the best for extracting the dynamic information of complex five-soliton molecules from TS-DFT data. The mean Pearson correlation coefficient (MPCC) between the predicted results and the real results is about 0.9975. Further, the ResNets in which the MPCC achieves 0.9906 also has the better ability of phase extraction than VGG which the MPCC is about 0.9739. The general applicability is demonstrated for extracting internal information from complex soliton molecule structures with high accuracy. The presented DCNs-based techniques can be employed to explore undiscovered mechanisms underlying the distribution and evolution of large numbers of solitons in dissipative systems in experimental research. |
| first_indexed | 2024-03-09T00:53:19Z |
| format | Article |
| id | doaj.art-31eafdbda109415db99d1f31ca5c7e94 |
| institution | Directory Open Access Journal |
| issn | 2304-6732 |
| language | English |
| last_indexed | 2024-03-09T00:53:19Z |
| publishDate | 2021-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Photonics |
| spelling | doaj.art-31eafdbda109415db99d1f31ca5c7e942023-12-11T17:01:13ZengMDPI AGPhotonics2304-67322021-02-01825110.3390/photonics8020051Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral InterferenceCaiyun Li0Jiangyong He1Yange Liu2Yang Yue3Luhe Zhang4Longfei Zhu5Mengjie Zhou6Congcong Liu7Kaiyan Zhu8Zhi Wang9Tianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaTianjin Key Laboratory of Optoelectronic Sensor and Sensing Network Technology, Institute of Modern Optics, Nankai University, Tianjin 300350, ChinaDeep neural networks have enabled the reconstruction of optical soliton molecules with more complex structures using the real-time spectral interferences obtained by photonic time-stretch dispersive Fourier transformation (TS-DFT) technology. In this paper, we propose to use three kinds of deep convolution networks (DCNs), including VGG, ResNets, and DenseNets, for revealing internal dynamics evolution of soliton molecules based on the real-time spectral interferences. When analyzing soliton molecules with equidistant composite structures, all three models are effective. The DenseNets with layers of 48 perform the best for extracting the dynamic information of complex five-soliton molecules from TS-DFT data. The mean Pearson correlation coefficient (MPCC) between the predicted results and the real results is about 0.9975. Further, the ResNets in which the MPCC achieves 0.9906 also has the better ability of phase extraction than VGG which the MPCC is about 0.9739. The general applicability is demonstrated for extracting internal information from complex soliton molecule structures with high accuracy. The presented DCNs-based techniques can be employed to explore undiscovered mechanisms underlying the distribution and evolution of large numbers of solitons in dissipative systems in experimental research.https://www.mdpi.com/2304-6732/8/2/51fiber nonlinearitiesdeep learning (DL)artificial intelligence (AI) |
| spellingShingle | Caiyun Li Jiangyong He Yange Liu Yang Yue Luhe Zhang Longfei Zhu Mengjie Zhou Congcong Liu Kaiyan Zhu Zhi Wang Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral Interference Photonics fiber nonlinearities deep learning (DL) artificial intelligence (AI) |
| title | Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral Interference |
| title_full | Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral Interference |
| title_fullStr | Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral Interference |
| title_full_unstemmed | Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral Interference |
| title_short | Comparing Performance of Deep Convolution Networks in Reconstructing Soliton Molecules Dynamics from Real-Time Spectral Interference |
| title_sort | comparing performance of deep convolution networks in reconstructing soliton molecules dynamics from real time spectral interference |
| topic | fiber nonlinearities deep learning (DL) artificial intelligence (AI) |
| url | https://www.mdpi.com/2304-6732/8/2/51 |
| work_keys_str_mv | AT caiyunli comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT jiangyonghe comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT yangeliu comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT yangyue comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT luhezhang comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT longfeizhu comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT mengjiezhou comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT congcongliu comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT kaiyanzhu comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference AT zhiwang comparingperformanceofdeepconvolutionnetworksinreconstructingsolitonmoleculesdynamicsfromrealtimespectralinterference |