TY - GEN
T1 - On the robustness of compressive sensing hyperspectral image reconstruction using convolutional neural network
AU - Gedalin, Daniel
AU - Heiser, Yaron
AU - Oiknine, Yaniv
AU - Stern, Adrian
N1 - Funding Information:
This research was supported by the Ministry of Science Technology & Space, Israel (grant 3-18410). Yaniv Oiknine wishes to thank Ministry of Science Technology & Space, Israel (grant 3-13351) for supporting this research.
Publisher Copyright:
© 2019 SPIE.
PY - 2019/1/1
Y1 - 2019/1/1
N2 - Hyperspectral imaging is applied in a wide range of defense, security and law enforcement applications. The spectral data caries valuable information for tasks such as identification, detection, and classification. However, the capturing of the spectral information, together with the spatial information, requires a significant acquisition effort. In the recent years we have developed several compressive hyperspectral imaging techniques demonstrating reduction of the captured data by at least an order of magnitude. However, compressive sensing techniques typically require computational heavy and time consuming iterative reconstruction algorithms. The computational burden is even more prominent in compressive spectral imaging due to the large amount of data involved. In this work we demonstrate the utilization of a convolutional neural network (CNN) for the reconstruction of spectral images captured with our Compressive Sensing-Miniature Ultraspectral Imager (CS-MUSI). We discuss the challenges of training the CNN for CS-MUSI and analyze the CNNbased reconstruction performance.
AB - Hyperspectral imaging is applied in a wide range of defense, security and law enforcement applications. The spectral data caries valuable information for tasks such as identification, detection, and classification. However, the capturing of the spectral information, together with the spatial information, requires a significant acquisition effort. In the recent years we have developed several compressive hyperspectral imaging techniques demonstrating reduction of the captured data by at least an order of magnitude. However, compressive sensing techniques typically require computational heavy and time consuming iterative reconstruction algorithms. The computational burden is even more prominent in compressive spectral imaging due to the large amount of data involved. In this work we demonstrate the utilization of a convolutional neural network (CNN) for the reconstruction of spectral images captured with our Compressive Sensing-Miniature Ultraspectral Imager (CS-MUSI). We discuss the challenges of training the CNN for CS-MUSI and analyze the CNNbased reconstruction performance.
KW - Compressive Sensing
KW - Deep Neural Networks
KW - Hyperspectral Reconstruction
KW - Inverse problem solving
UR - http://www.scopus.com/inward/record.url?scp=85077333696&partnerID=8YFLogxK
U2 - 10.1117/12.2533113
DO - 10.1117/12.2533113
M3 - Conference contribution
AN - SCOPUS:85077333696
T3 - Proceedings of SPIE - The International Society for Optical Engineering
BT - Artificial Intelligence and Machine Learning in Defense Applications
A2 - Dijk, Judith
PB - SPIE
T2 - Artificial Intelligence and Machine Learning in Defense Applications 2019
Y2 - 10 September 2019 through 12 September 2019
ER -