Ghost Imaging Quality Optimization Based on Deep Convolutional Generative Adversarial Networks

Maoxin Hou; Zhaotao Liu

doi:10.3788/LOP232421

Laser & Optoelectronics Progress, Volume. 61, Issue 10, 1011010(2024)

Ghost Imaging Quality Optimization Based on Deep Convolutional Generative Adversarial Networks

Maoxin Hou^1、* and Zhaotao Liu²

¹Collective Intelligence & Collaboration Laboratory, Zhongbing Intelligent Innovation Research Institute Limited Liabilty Company, Beijing 100072, China

²China North Vehicle Research Institute, Beijing 100072, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(12)

Fig. 1. Schematic of computational ghost imaging based on Hadamard speckle

Download full size

Fig. 2. SDCGAN-GI architecture

Download full size

Fig. 3. SDCGAN-GI architecture model

Download full size

Fig. 4. Handwritten digital images reconstructed at different sampling rates

Download full size

Fig. 5. Comparison of the results of different algorithms at a sampling rate of 0.0625

Download full size

Fig. 6. Comparison of the results of different algorithms at a sampling rate of 0.0156

Download full size

Fig. 7. PSNR at different sampling rates

Download full size

Fig. 8. Comparison of reconstruction results of SDCGAN and u-net at a sampling rate of 0.25

Download full size

Fig. 9. Comparison of reconstruction results of SDCGAN and u-net at a sampling rate of 0.0625

Download full size

Fig. 10. Average peak signal-to-noise ratio and average structural similarity between SDCGAN and u-net at different sampling rates

Download full size

Table 1. Comparison of average peak signal-to-noise ratio between SDCGAN-GI and u-net-GI at different sampling rates
View table
Table 1. Comparison of average peak signal-to-noise ratio between SDCGAN-GI and u-net-GI at different sampling rates
Sampling rate /% SDCGAN-GI u-net-GI
0.39 9.8157 9.4954
1.56 11.6554 10.2425
6.25 12.5023 10.5131
25 18.0551 13.0557

Table 2. Comparison of average structural similarity between SDCGAN-GI and u-net-GI at different sampling rates
View table
Table 2. Comparison of average structural similarity between SDCGAN-GI and u-net-GI at different sampling rates
Sampling rate /% SDCGAN-GI u-net-GI
0.39 0.1187 0.0779
1.56 0.4082 0.2587
6.25 0.4554 0.2998
25 0.7630 0.5360

Tools

Get Citation

Copy Citation Text

Maoxin Hou, Zhaotao Liu. Ghost Imaging Quality Optimization Based on Deep Convolutional Generative Adversarial Networks[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1011010

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Imaging Systems

Received: Nov. 2, 2023

Accepted: Dec. 26, 2023

Published Online: Apr. 29, 2024

The Author Email: Hou Maoxin (wang17835132895@163.com)

DOI:10.3788/LOP232421

CSTR:32186.14.LOP232421

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Comparison of average peak signal-to-noise ratio between SDCGAN-GI and u-net-GI at different sampling rates

Table 1. Comparison of average peak signal-to-noise ratio between SDCGAN-GI and u-net-GI at different sampling rates

Table 2. Comparison of average structural similarity between SDCGAN-GI and u-net-GI at different sampling rates

Table 2. Comparison of average structural similarity between SDCGAN-GI and u-net-GI at different sampling rates