Super-Resolution Restoration of Low-Resolution THz Camera Images

Damyanov, D.; Friederich, B.; Kohl, J. L.; Liu, X.; Kolpatzeck, Kevin; Schultze, Thorsten; Czylwik, Andreas; Balzer, Jan C.; Willms, Ingolf

doi:10.1109/iwmts.2019.8823667

Cited by 2 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…also used POCS and IBP to reconstruct high‐resolution THz images from four THz TDS images [15]; Damyanov et al. combined the classical super‐resolution method with Wiener filter compensation system and applied it to the terahertz imaging system [16]. These methods solve the problem of low‐resolution of the terahertz imaging systems to some extent, but the performance improvement of traditional methods for super‐resolution tasks is limited.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

UHA‐CycleGAN: Unpaired hybrid attention network based on CycleGAN for terahertz image super‐resolution

Liu

Guo

Liu

2023

IET Image Processing

View full text Add to dashboard Cite

In recent years, terahertz imaging technology has been widely used in security, medicine, and other fields. However, the image resolution is low due to the limits of imaging equipment and diffraction. Traditional super-resolution methods based on machine learning use artificial paired image datasets, and their image degradation process is quite different from the actual terahertz imaging mechanism. In the paper, an unpaired hybrid attention network is proposed, using the real unpaired high-resolution and low-resolution terahertz images as the original input, two sub-networks are trained: degenerate reconstruction network and super-resolution network. The degenerate reconstruction network is used to learn the degradation process of the real terahertz imaging system, and reconstruct the generated degradation images into the fixed mapping down-sampled image of highresolution terahertz image. The super-resolution network uses the generated paired image with fixed correspondence to train the super-resolution network of paired terahertz images, to improve the imaging resolution of the network in the super-resolution work of real low-resolution terahertz images. In addition, the no-reference image quality evaluation system is introduced to objectively evaluate the network performance. The experimental results show that the UHA-CycleGAN network outperforms the traditional paired super-resolution network on both open-source and self-built datasets.

show abstract

Section: Related Workmentioning

confidence: 99%

“…See Table 1 for specific super parameter settings and weight settings in loss function. In the experiment, the batchsize is set to 16. All experiments are trained on a RTX3090 graphics card using pytorch.…”

Section: Network Configurationmentioning

confidence: 99%