Regular Articles

Full Text / References

Trial of Deep Learning for Image Reconstruction of Lens-Less Microwave Holography

Ryo MANABE, Hayato TSUCHIYA¹⁾ and Mayuko KOGA

University of Hyogo, 2167 Shosha, Himeji, Hyogo 671-2280, Japan

1)

System Technology Development Center, Kawasaki Heavy Industries, Ltd., 1-1 Kawasaki-cho, Akashi, Hyogo 673-8666, Japan

(Received 26 December 2021 / Accepted 24 April 2022 / Published 22 June 2022)

Abstract

We perform the principal verification of reconstructing object surface images by using deep learning. Using the deep learning neural network based on convolutional neural networks, simple object surface images with 128 × 128 pixels are reasonably reconstructed with up-converting from rough microwave signal images with 16 × 16 pixels. The model captures large structural features of the object surface images even with small number of training data. As the number of training data increases, it captures small structures of objects. It is also found that noises of input signal images affect reconstructions of small structures of objects.

Copyright (c) 2022 The Japan Society of Plasma Science and Nuclear Fusion Research

Keywords

microwave holography, deep learning, convolutional neural networks (CNNs), lens-less imaging, image reconstruction

DOI: 10.1585/pfr.17.2401072

Full Text

PDF format (4.0Mbytes)

References

• [1] T. Yamada et al., Nature Phys. 4, 721 (2008).
• [2] S-I. Itoh, Plasma Fusion Res. 4, 038 (2009).
• [3] T. Munsat, E. Mazzucato and H. Park, Rev. Sci. Instrum. 74, 1426 (2003).
• [4] S. Yamaguchi et al., Rev. Sci. Instrum. 77, 10E930 (2006).
• [5] W. Lee et al., JINST 7, C01070 (2012).
• [6] M. Muscatello et al., Rev. Sci. Instrum. 85, 11D702 (2014).
• [7] H. Tsuchiya et al., Plasma Fusion Res. 13, 3402063 (2018).
• [8] B. Sun et al., Science 340, 844 (2013).
• [9] J. Wu et al., Light: Science & Applications 9, 53 (2020).
• [10] H. Tsuchiya et al., Plasma Fusion Res. 14, 3402146 (2019).
• [11] M. Koga et al., Plasma Fusion Res. 16, 1402063 (2021).
• [12] Y. Nagayama et al., Rev Sci. Instrum. 88, 044703 (2017).