EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Sensor-Physics-Driven Noise Modeling for Low-Light Imaging Using Adversarial Learning

Peihua Zhao, Baopeng Li, Hui Zhao, Wansha Wen, Wei Gao, Xuewu Fan

Applied Sciences · 18.03.2026 · 16 (6) · 2948

DOI: 10.3390/app16062948 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 18.03.2026. Verfasst von Peihua Zhao und 5 weiteren Autor:innen. Die Arbeit nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Western Young Scholars of the Chinese Academy of Sciences.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

High-fidelity imaging in extreme low light is challenged by ultra-low signal-to-noise ratios. We propose a hybrid noise modeling framework integrating physical priors with generative adversarial networks (GANs). The method simulates photon shot noise via Poisson distribution and incorporates readout, row, and quantization noise. A multi-layer perceptron (MLP) dynamically maps ISO levels to noise intensities in logarithmic space, followed by a residual U-Net for non-linear refinement. Results on the SID datasets show that our method outperforms state-of-the-art approaches in terms of Average Kullback–Leibler Divergence (AKLD). Denoising networks trained on our synthetic noise achieve performances comparable to those trained on real-world paired datasets.

Förderung

  • National Natural Science Foundation of China
  • Western Young Scholars of the Chinese Academy of Sciences

Zitieren

APA

Zhao, P., Li, B., Zhao, H., Wen, W., Gao, W., & Fan, X. (2026). Sensor-Physics-Driven Noise Modeling for Low-Light Imaging Using Adversarial Learning. Applied Sciences, 16(6), 2948. https://doi.org/10.3390/app16062948

Vancouver

Zhao P, Li B, Zhao H, Wen W, Gao W, Fan X. Sensor-Physics-Driven Noise Modeling for Low-Light Imaging Using Adversarial Learning. Applied Sciences. 2026;16(6):2948. doi:10.3390/app16062948

BibTeX
@article{Zhao2026,
  author = {Zhao, Peihua and Li, Baopeng and Zhao, Hui and Wen, Wansha and Gao, Wei and Fan, Xuewu},
  title = {Sensor-Physics-Driven Noise Modeling for Low-Light Imaging Using Adversarial Learning},
  journal = {Applied Sciences},
  year = {2026},
  volume = {16},
  number = {6},
  pages = {2948},
  doi = {10.3390/app16062948}
}
Kostenlos starten →