Compressive Sensing Spectroscopy Using a Residual Convolutional Neural Network
Compressive sensing (CS) spectroscopy is well known for developing a compact spectrometer which consists of two parts: compressively measuring an input spectrum and recovering the spectrum using reconstruction techniques. Our goal here is to propose a novel residual convolutional neural network (Res...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-01-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/3/594 |
| _version_ | 1811307370128605184 |
|---|---|
| author | Cheolsun Kim Dongju Park Heung-No Lee |
| author_facet | Cheolsun Kim Dongju Park Heung-No Lee |
| author_sort | Cheolsun Kim |
| collection | DOAJ |
| description | Compressive sensing (CS) spectroscopy is well known for developing a compact spectrometer which consists of two parts: compressively measuring an input spectrum and recovering the spectrum using reconstruction techniques. Our goal here is to propose a novel residual convolutional neural network (ResCNN) for reconstructing the spectrum from the compressed measurements. The proposed ResCNN comprises learnable layers and a residual connection between the input and the output of these learnable layers. The ResCNN is trained using both synthetic and measured spectral datasets. The results demonstrate that ResCNN shows better spectral recovery performance in terms of average root mean squared errors (RMSEs) and peak signal to noise ratios (PSNRs) than existing approaches such as the sparse recovery methods and the spectral recovery using CNN. Unlike sparse recovery methods, ResCNN does not require <i>a priori</i> knowledge of a sparsifying basis nor prior information on the spectral features of the dataset. Moreover, ResCNN produces stable reconstructions under noisy conditions. Finally, ResCNN is converged faster than CNN. |
| first_indexed | 2024-04-13T09:03:12Z |
| format | Article |
| id | doaj.art-529fa0090e4e4060b52a111c31d66fca |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-13T09:03:12Z |
| publishDate | 2020-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-529fa0090e4e4060b52a111c31d66fca2022-12-22T02:53:03ZengMDPI AGSensors1424-82202020-01-0120359410.3390/s20030594s20030594Compressive Sensing Spectroscopy Using a Residual Convolutional Neural NetworkCheolsun Kim0Dongju Park1Heung-No Lee2School of Electrical Engineering and Computer Science, Gwangju Institute of Science and Technology, Gwangju 61005, KoreaSchool of Electrical Engineering and Computer Science, Gwangju Institute of Science and Technology, Gwangju 61005, KoreaSchool of Electrical Engineering and Computer Science, Gwangju Institute of Science and Technology, Gwangju 61005, KoreaCompressive sensing (CS) spectroscopy is well known for developing a compact spectrometer which consists of two parts: compressively measuring an input spectrum and recovering the spectrum using reconstruction techniques. Our goal here is to propose a novel residual convolutional neural network (ResCNN) for reconstructing the spectrum from the compressed measurements. The proposed ResCNN comprises learnable layers and a residual connection between the input and the output of these learnable layers. The ResCNN is trained using both synthetic and measured spectral datasets. The results demonstrate that ResCNN shows better spectral recovery performance in terms of average root mean squared errors (RMSEs) and peak signal to noise ratios (PSNRs) than existing approaches such as the sparse recovery methods and the spectral recovery using CNN. Unlike sparse recovery methods, ResCNN does not require <i>a priori</i> knowledge of a sparsifying basis nor prior information on the spectral features of the dataset. Moreover, ResCNN produces stable reconstructions under noisy conditions. Finally, ResCNN is converged faster than CNN.https://www.mdpi.com/1424-8220/20/3/594spectroscopycompressed sensingdeep learninginverse problemssparse recoverydictionary learning |
| spellingShingle | Cheolsun Kim Dongju Park Heung-No Lee Compressive Sensing Spectroscopy Using a Residual Convolutional Neural Network Sensors spectroscopy compressed sensing deep learning inverse problems sparse recovery dictionary learning |
| title | Compressive Sensing Spectroscopy Using a Residual Convolutional Neural Network |
| title_full | Compressive Sensing Spectroscopy Using a Residual Convolutional Neural Network |
| title_fullStr | Compressive Sensing Spectroscopy Using a Residual Convolutional Neural Network |
| title_full_unstemmed | Compressive Sensing Spectroscopy Using a Residual Convolutional Neural Network |
| title_short | Compressive Sensing Spectroscopy Using a Residual Convolutional Neural Network |
| title_sort | compressive sensing spectroscopy using a residual convolutional neural network |
| topic | spectroscopy compressed sensing deep learning inverse problems sparse recovery dictionary learning |
| url | https://www.mdpi.com/1424-8220/20/3/594 |
| work_keys_str_mv | AT cheolsunkim compressivesensingspectroscopyusingaresidualconvolutionalneuralnetwork AT dongjupark compressivesensingspectroscopyusingaresidualconvolutionalneuralnetwork AT heungnolee compressivesensingspectroscopyusingaresidualconvolutionalneuralnetwork |