EUVIMEDEUVIMED entdecken
Crossref

DestripeCycleGAN: Stripe Simulation CycleGAN for Unsupervised Infrared Image Destriping

Shiqi Yang, Hanlin Qin, Shuai Yuan, Xiang Yan, Hossein Rahmani

IEEE Transactions on Instrumentation and Measurement · 2024 · 73 · 1-14

DOI: 10.1109/tim.2024.3476560 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Instrumentation and Measurement, 2024. Verfasst von Shiqi Yang und 4 weiteren Autor:innen. Die Arbeit wurde bisher 29-mal zitiert (Crossref) und nennt 69 Literaturangaben. Gefördert von: Shaanxi Province Key Research and Development Plan Project, Shaanxi Province Science and Technology Plan Project, Xian City Science and Technology Plan Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

CycleGAN has been proved to be an advanced approach for unsupervised image restoration. This framework consists of two generators: a denoising one for inference and an auxiliary one for modeling noise to fulfill cycle-consistent constraints. However, when applied to the infrared destriping task, it becomes challenging for the vanilla auxiliary generator to consistently produce vertical noise under unsupervised constraints. This poses a threat to the effectiveness of the cycle-consistent loss, leading to stripe noise residual in the denoised image. To address the above issue, we present a novel framework for single-frame infrared image destriping, named DestripeCycleGAN. In this model, the conventional auxiliary generator is replaced with an a priori stripe generation model (SGM) to introduce vertical stripe noise in the clean data, and the gradient map is employed to re-establish cycle consistency. Meanwhile, a Haar wavelet background guidance module (HBGM) has been designed to minimize the divergence of background details between the different domains. To preserve vertical edges, a multilevel wavelet U-Net (MWUNet) is proposed as the denoising generator, which utilizes the Haar wavelet transform as the sampler to decline directional information loss. Moreover, it incorporates the group fusion block (GFB) into skip connections to fuse the multiscale features and build the context of long-distance dependencies. Extensive experiments on real and synthetic data demonstrate that our DestripeCycleGAN surpasses the state-of-the-art methods in terms of visual quality and quantitative evaluation. Our code is available at https://github.com/xdFai/DestripeCycleGAN.

Förderung

  • Shaanxi Province Key Research and Development Plan Project
  • Shaanxi Province Science and Technology Plan Project
  • Xian City Science and Technology Plan Project
  • Aeronautical Science Foundation of China
  • Natural Science Foundation Explore of Zhejiang province
  • Natural Science Foundation of Ningbo Municipality
  • Technology Area Foundation of China
  • China Scholarship Council

Zitieren

APA

Yang, S., Qin, H., Yuan, S., Yan, X., & Rahmani, H. (2024). DestripeCycleGAN: Stripe Simulation CycleGAN for Unsupervised Infrared Image Destriping. IEEE Transactions on Instrumentation and Measurement, 73, 1-14. https://doi.org/10.1109/tim.2024.3476560

Vancouver

Yang S, Qin H, Yuan S, Yan X, Rahmani H. DestripeCycleGAN: Stripe Simulation CycleGAN for Unsupervised Infrared Image Destriping. IEEE Transactions on Instrumentation and Measurement. 2024;73:1-14. doi:10.1109/tim.2024.3476560

BibTeX
@article{Yang2024,
  author = {Yang, Shiqi and Qin, Hanlin and Yuan, Shuai and Yan, Xiang and Rahmani, Hossein},
  title = {DestripeCycleGAN: Stripe Simulation CycleGAN for Unsupervised Infrared Image Destriping},
  journal = {IEEE Transactions on Instrumentation and Measurement},
  year = {2024},
  volume = {73},
  pages = {1--14},
  doi = {10.1109/tim.2024.3476560}
}
Kostenlos starten →