全息术
电子全息术
稳健性(进化)
光学
相(物质)
材料科学
噪音(视频)
降噪
相位恢复
高斯分布
散粒噪声
电子
边带
计算机科学
数字全息术
高斯噪声
物理
高斯过程
航程(航空)
过程(计算)
还原(数学)
纳米尺度
作者
Shuhang Liao,Jiaqi Su,Binye Liang,Zi‐An Li
出处
期刊:Physica Scripta
[IOP Publishing]
日期:2026-09-16
卷期号:101 (39): 395903-395903
标识
DOI:10.1088/1402-4896/aea8e4
摘要
Abstract Low-dose electron holography enables quantitative phase imaging of electron-beam-sensitive materials and weak electromagnetic fields, but reduced electron counts introduce severe noise that limits phase reconstruction accuracy and stability. Here, we develop a denoising method for off-axis electron holography based on frequency-domain component separation and Gaussian process regression (GPR). By selectively preserving sideband information and applying GPR denoising to the non-sideband component, the proposed approach suppresses noise-induced fluctuations while maintaining holographic phase information. The method is validated using simulated holograms over a wide range of signal-to-noise ratios and low-dose experimental electron holograms. Compared with conventional filtering methods and wavelet-based approaches, GPR provides improved noise suppression, reduced phase reconstruction errors, and enhanced robustness under extremely low-dose conditions. This approach offers a practical strategy for improving quantitative phase measurements in low-dose electron holography and broadens the applicability of electron holography for sensitive nanoscale materials and electromagnetic-field characterization.
科研通智能强力驱动
Strongly Powered by AbleSci AI