Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-Apertures
This work aims to simplify the characterization process of coded-apertures for computational imaging (CI) at microwave frequencies. A major benefit of the presented technique is the minimization of the processing time needed to calculate the system sensing matrix for microwave CI-based compressive s...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10415384/ |
| _version_ | 1827357596698279936 |
|---|---|
| author | Jiaming Zhang Rahul Sharma Maria Garcia-Fernandez Guillermo Alvarez-Narciandi Muhammad Ali Babar Abbasi Okan Yurduseven |
| author_facet | Jiaming Zhang Rahul Sharma Maria Garcia-Fernandez Guillermo Alvarez-Narciandi Muhammad Ali Babar Abbasi Okan Yurduseven |
| author_sort | Jiaming Zhang |
| collection | DOAJ |
| description | This work aims to simplify the characterization process of coded-apertures for computational imaging (CI) at microwave frequencies. A major benefit of the presented technique is the minimization of the processing time needed to calculate the system sensing matrix for microwave CI-based compressive sensing applications. To achieve this, a deep learning-based approach which is capable of generating the sensing matrix using features learned directly from the coded-aperture distribution is proposed. To avoid the vanishing gradient problem, the proposed deep learning network contains skip connections. Using a dataset of 1,000 testing samples, the average normalized mean-squared-error (NMSE) calculated between the sensing matrix generated by the conventional method and that predicted by the proposed network is 0.0036. Moreover, the average mean-squared-error (MSE) calculated between the images reconstructed using the conventional and the predicted sensing matrix is 0.00297. In addition to providing high-fidelity estimations with minimized error, we demonstrate that using the trained network, the prediction of the sensing matrix can be achieved in 0.212 s, corresponding to a 65% reduction in the computation time needed to calculate the sensing matrix. This has significant outcomes in achieving real-time operation of CI-based microwave imaging systems. |
| first_indexed | 2024-03-08T05:35:10Z |
| format | Article |
| id | doaj.art-a5b463b1534743f79d9b0bafdbe430be |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-08T05:35:10Z |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-a5b463b1534743f79d9b0bafdbe430be2024-02-06T00:01:29ZengIEEEIEEE Access2169-35362024-01-0112168441685510.1109/ACCESS.2024.335943510415384Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-AperturesJiaming Zhang0https://orcid.org/0009-0004-2658-2851Rahul Sharma1https://orcid.org/0000-0002-1177-1156Maria Garcia-Fernandez2https://orcid.org/0000-0001-8935-1912Guillermo Alvarez-Narciandi3https://orcid.org/0000-0001-9286-4372Muhammad Ali Babar Abbasi4https://orcid.org/0000-0002-1283-4614Okan Yurduseven5https://orcid.org/0000-0002-0242-3029Centre for Wireless Innovation (CWI), School of Electronics, Electrical Engineering and Computer Science (EEECS), Queen’s University Belfast, Belfast, U.KCentre for Wireless Innovation (CWI), School of Electronics, Electrical Engineering and Computer Science (EEECS), Queen’s University Belfast, Belfast, U.KCentre for Wireless Innovation (CWI), School of Electronics, Electrical Engineering and Computer Science (EEECS), Queen’s University Belfast, Belfast, U.KCentre for Wireless Innovation (CWI), School of Electronics, Electrical Engineering and Computer Science (EEECS), Queen’s University Belfast, Belfast, U.KCentre for Wireless Innovation (CWI), School of Electronics, Electrical Engineering and Computer Science (EEECS), Queen’s University Belfast, Belfast, U.KCentre for Wireless Innovation (CWI), School of Electronics, Electrical Engineering and Computer Science (EEECS), Queen’s University Belfast, Belfast, U.KThis work aims to simplify the characterization process of coded-apertures for computational imaging (CI) at microwave frequencies. A major benefit of the presented technique is the minimization of the processing time needed to calculate the system sensing matrix for microwave CI-based compressive sensing applications. To achieve this, a deep learning-based approach which is capable of generating the sensing matrix using features learned directly from the coded-aperture distribution is proposed. To avoid the vanishing gradient problem, the proposed deep learning network contains skip connections. Using a dataset of 1,000 testing samples, the average normalized mean-squared-error (NMSE) calculated between the sensing matrix generated by the conventional method and that predicted by the proposed network is 0.0036. Moreover, the average mean-squared-error (MSE) calculated between the images reconstructed using the conventional and the predicted sensing matrix is 0.00297. In addition to providing high-fidelity estimations with minimized error, we demonstrate that using the trained network, the prediction of the sensing matrix can be achieved in 0.212 s, corresponding to a 65% reduction in the computation time needed to calculate the sensing matrix. This has significant outcomes in achieving real-time operation of CI-based microwave imaging systems.https://ieeexplore.ieee.org/document/10415384/Computational imagingdeep learningimage reconstructionmicrowave imagingsensing matrix |
| spellingShingle | Jiaming Zhang Rahul Sharma Maria Garcia-Fernandez Guillermo Alvarez-Narciandi Muhammad Ali Babar Abbasi Okan Yurduseven Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-Apertures IEEE Access Computational imaging deep learning image reconstruction microwave imaging sensing matrix |
| title | Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-Apertures |
| title_full | Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-Apertures |
| title_fullStr | Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-Apertures |
| title_full_unstemmed | Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-Apertures |
| title_short | Deep Learning for Sensing Matrix Prediction in Computational Microwave Imaging With Coded-Apertures |
| title_sort | deep learning for sensing matrix prediction in computational microwave imaging with coded apertures |
| topic | Computational imaging deep learning image reconstruction microwave imaging sensing matrix |
| url | https://ieeexplore.ieee.org/document/10415384/ |
| work_keys_str_mv | AT jiamingzhang deeplearningforsensingmatrixpredictionincomputationalmicrowaveimagingwithcodedapertures AT rahulsharma deeplearningforsensingmatrixpredictionincomputationalmicrowaveimagingwithcodedapertures AT mariagarciafernandez deeplearningforsensingmatrixpredictionincomputationalmicrowaveimagingwithcodedapertures AT guillermoalvareznarciandi deeplearningforsensingmatrixpredictionincomputationalmicrowaveimagingwithcodedapertures AT muhammadalibabarabbasi deeplearningforsensingmatrixpredictionincomputationalmicrowaveimagingwithcodedapertures AT okanyurduseven deeplearningforsensingmatrixpredictionincomputationalmicrowaveimagingwithcodedapertures |