Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Davis, Allison, Shen, Yezhi, Ji, Xiaoyu, Zhu, Fengqing
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2510.03452
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916987991490560
author Davis, Allison
Shen, Yezhi
Ji, Xiaoyu
Zhu, Fengqing
author_facet Davis, Allison
Shen, Yezhi
Ji, Xiaoyu
Zhu, Fengqing
contents Structured illumination (SI) enhances image resolution and contrast by projecting patterned light onto a sample. In two-phase optical-sectioning SI (OS-SI), reduced acquisition time introduces residual artifacts that conventional denoising struggles to suppress. Deep learning offers an alternative to traditional methods; however, supervised training is limited by the lack of clean, optically sectioned ground-truth data. We investigate encoder-decoder networks for artifact reduction in two-phase OS-SI, using synthetic training pairs formed by applying real artifact fields to synthetic images. An asymmetrical denoising autoencoder (DAE) and a U-Net are trained on the synthetic data, then evaluated on real OS-SI images. Both networks improve image clarity, with each excelling against different artifact types. These results demonstrate that synthetic training enables supervised denoising of OS-SI images and highlight the potential of encoder-decoder networks to streamline reconstruction workflows.
format Preprint
id arxiv_https___arxiv_org_abs_2510_03452
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Denoising of Two-Phase Optically Sectioned Structured Illumination Reconstructions Using Encoder-Decoder Networks
Davis, Allison
Shen, Yezhi
Ji, Xiaoyu
Zhu, Fengqing
Computer Vision and Pattern Recognition
Structured illumination (SI) enhances image resolution and contrast by projecting patterned light onto a sample. In two-phase optical-sectioning SI (OS-SI), reduced acquisition time introduces residual artifacts that conventional denoising struggles to suppress. Deep learning offers an alternative to traditional methods; however, supervised training is limited by the lack of clean, optically sectioned ground-truth data. We investigate encoder-decoder networks for artifact reduction in two-phase OS-SI, using synthetic training pairs formed by applying real artifact fields to synthetic images. An asymmetrical denoising autoencoder (DAE) and a U-Net are trained on the synthetic data, then evaluated on real OS-SI images. Both networks improve image clarity, with each excelling against different artifact types. These results demonstrate that synthetic training enables supervised denoising of OS-SI images and highlight the potential of encoder-decoder networks to streamline reconstruction workflows.
title Denoising of Two-Phase Optically Sectioned Structured Illumination Reconstructions Using Encoder-Decoder Networks
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.03452