Quantitatively mapping local quality of super-resolution microscopy by rolling Fourier ring correlation

Abstract

In fluorescence microscopy, computational algorithms have been developed to suppress noise, enhance contrast, and even enable super-resolution (SR). However, the local quality of the images may vary on multiple scales, and these differences can lead to misconceptions, which is especially intractable in emerging deep-learning ones. Current mapping methods fail to finely estimate the local quality, challenging to associate the SR scale content. Here, we develop a rolling Fourier ring correlation (rFRC) framework to evaluate the reconstruction uncertainties down to SR scale. To visually pinpoint regions with low reliability, a filtered rFRC is combined with a modified resolution scaled error map (RSM), offering a comprehensive and concise map for further examination. We demonstrate their performances on various SR imaging modalities, and the resulting quantitative maps enable better SR images integrated from different reconstructions. Beyond that, we provide a strategy for learning-based restorations, allowing a direct detection of both data and model uncertainties, and expect the representative cases can inspire further advances in this rapidly developing field.

Time-lapse Image Super-resolution Neural Network with Reliable Confidence Quantification for Optical Microscopy

“…Fig. 3d shows the representative SR reconstruction, corresponding absolute error map, and confidence maps generated by Bayesian DPA-TISR after correction and rolling Fourier ring correlation (rFRC) 32 .…”

“… a , b , Reliability diagrams generated by Bayesian DPA-TISR models before (left panel) and after (right panel) confidence calibration for microtubule (a) and mitochondrion images (b). c , Representative wide-field (WF) images (bottom left corner of the first column), TISR images (top right corner of the first column), absolute error maps (second column), rFRC maps generated with rolling Fourier ring correlation 32 (third column), and confidence map estimated by Bayesian DPA-TISR models (four column) of microtubules and mitochondria. Scale bar: 1μm (c), and 0.5μm (zoom-in regions in c).…”

Dark-based Optical Sectioning assists Background Removal in Fluorescence Microscopy

“…4(i). Although Dark sectioning cannot significantly improve the resolution on the basis of Confocal (resolution of Dark confocal being 463.831nm, Confocal being 446.679 nm using rFRC 37 ). Dark sectioning can strengthen the in-focus information to provide more information for SACD, so the resolution of SCAD when imaging thick samples like pine young staminate has been significantly improved from 353.314nm to 216.784nm using rFRC with a successfully SACD reconstruction (RSE of Dark SACD being 0.0208, RSE of SACD being 0.0210) under the same parameters.…”

“…We use rFRC mapping 37 to analysis the resolution of SOFI/SACD. The rFRC map denotes the spatial distribution of resolution.…”

Squaraine Dyes Exhibit Spontaneous Fluorescence Blinking That Enables Live-Cell Nanoscopy

“…Fourier ring correlation (FRC) analysis demonstrated that the reconstructed image has a resolution of approximately 69 nm (Figure 4g). 79 Analysis of rolling Fourier ring correlation 80 (rFRC) on reconstructed images reveals varying local resolutions with a median value of 43 nm (Figure S20), potentially attributed to differing focal planes of the inner membrane and pseudopodia, probe signal strength variations, and membrane fluidity affecting localization accuracy in live cells. In brief, SQ-F-based super-resolution imaging of living cell membranes visualized the ultrastructure of pseudopodia.…”