Figure - PMC

Skip to main content

An official website of the United States government

Here's how you know

Here's how you know

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

View full-text article in PMC

. 2022 Oct 31;19(11):1427–1437. doi: 10.1038/s41592-022-01652-7

Search in PMC
Search in PubMed
View in NLM Catalog
Add to search

© This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2022

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

PMC Copyright notice

Fig. 4 — a, Lateral and axial planes from images of a fixed U2OS cell stained with Alexa Fluor 568 Phalloidin, comparing widefield raw data, RLD with 100 iterations, Thunder output, RLN result and the registered confocal data as a ground truth reference. RLN was trained with synthetic mixed structures. RLN predictions show better restoration than RLD and Thunder, particularly along the axial dimension. PSNR and SSIM analysis using the confocal data as ground truth confirm this result (raw widefield, SSIM 0.57 ± 0.03; PSNR 28.7 ± 0.9; RLD, SSIM 0.67 ± 0.03, PSNR 30.0 ± 0.7; Thunder, SSIM 0.63 ± 0.05, PSNR 30.0 ± 0.7; RLN, SSIM 0.73 ± 0.02, PSNR 30.9 ± 0.9, n = 4 volumes). b, Lateral and axial planes of images of nuclei pore complexes in a fixed COS-7 cell immunolabeled with primary mouse anti-Nup clone Mab414 and goat-antimouse IgG secondary antibody conjugated with Star635P, comparing widefield input, RLD with 100 iterations, Thunder output, RLN prediction and registered confocal data as a ground truth reference. c, Magnified views of the blue rectangle in b. d, Line profiles across the red and magenta lines in the lateral and axial views in b. RLN was trained with synthetic mixed structures. Both visual analysis (for example, red arrows) and line intensity profiles demonstrate that RLN restoration outperform Thunder (obvious artifacts indicated by orange arrows) and RLD in both lateral and axial views, showing detail that approaches the confocal reference. PSNR and SSIM analysis using the registered confocal results as ground truth confirm this result (raw widefield input SSIM 0.78 ± 0.04, PSNR 34.5 ± 0.4; RLD SSIM 0.79 ± 0.02, PSNR 36.7 ± 0.5; Thunder SSIM 0.80 ± 0.04, PSNR 36.7 ± 0.4; RLN SSIM 0.86 ± 0.01, PSNR 37.5 ± 0.3, n = 4 volumes). Scale bars, a 10 μm; b 10 μm; c 3 μm. Experiments repeated four times for both a and b, representative data from single experiment are shown.