DPGI: A Two-Stage Framework for Robust Super-Resolution Ghost Imaging in Noisy Environments.

Saved in:

Bibliographic Details
Title:	DPGI: A Two-Stage Framework for Robust Super-Resolution Ghost Imaging in Noisy Environments.
Authors:	Chen, Zhongyang¹ czy112@xynu.edu.cn, Xuan, Hejun² xuanhejun0896@xynu.edu.cn, Yang, Zhipeng¹ yangzp@xynu.edu.cn
Source:	IAENG International Journal of Computer Science. Jul2026, Vol. 53 Issue 7, p2575-2582. 8p.
Subjects:	Image reconstruction, Mathematical regularization, Deep learning, Speckle interference, Robust statistics, High resolution imaging
Abstract:	Ghost imaging (GI) faces a fundamental tradeoff between acquisition speed and reconstruction fidelity, particularly under low sampling rates and high environmental noise. To address this, we propose Deep Processing Ghost Imaging (DPGI), a robust two-stage framework. The method first employs a Gaussian regularization step to transform chaotic speckle noise into a tractable structured blur, creating a clean prior. Subsequently, a Deep Plug-and-Play Super-Resolution (DPSR) model leverages this known degradation kernel to precisely restore high-frequency details. Comprehensive simulations demonstrate that DPGI consistently outperforms traditional GI, BM3D, and deep learning benchmarks. Specifically, under challenging conditions, DPGI achieves performance gains of up to 18.4% in PSNR and 45.1% in SSIM compared to end-to-end approaches, proving its superior robustness in noisy environments. [ABSTRACT FROM AUTHOR]
	Copyright of IAENG International Journal of Computer Science is the property of International Association of Engineers (IAENG) and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database:	Engineering Source

Description
Abstract:	Ghost imaging (GI) faces a fundamental tradeoff between acquisition speed and reconstruction fidelity, particularly under low sampling rates and high environmental noise. To address this, we propose Deep Processing Ghost Imaging (DPGI), a robust two-stage framework. The method first employs a Gaussian regularization step to transform chaotic speckle noise into a tractable structured blur, creating a clean prior. Subsequently, a Deep Plug-and-Play Super-Resolution (DPSR) model leverages this known degradation kernel to precisely restore high-frequency details. Comprehensive simulations demonstrate that DPGI consistently outperforms traditional GI, BM3D, and deep learning benchmarks. Specifically, under challenging conditions, DPGI achieves performance gains of up to 18.4% in PSNR and 45.1% in SSIM compared to end-to-end approaches, proving its superior robustness in noisy environments. [ABSTRACT FROM AUTHOR]
ISSN:	1819656X