FlexSIM
November 29, 2025 · View on GitHub
FlexSIM, for flexible SIM reconstruction, aims at providing reliable SIM reconstructions for a variety of SIM data, going from ``ideal'' ones acquired under standardized protocols and configurations, to ones obtained under more challenging settings and subject to reconstruction artifacts.
More details can be found in the following paper:
Surpassing Light Inhomogeneities in Structured-Illumination Microscopy with FlexSIM.,
Journal of Microscopy (2024),
E. Soubies, A. Nogueron, F. Pelletier, T. Mangeat, C. Leterrier, M. Unser, and D. Sage.
Outline
News
- May 2025: New saving options and integration of the newly proposed Dark-Sectioning method.
- Feb 2025: New examples have been added (see below)
- Dec 2024: The handling of temporal stacks has been improved. A new parameter eqOrr has been introduced to impose a soft constraint of equally spaced orientations during patterns estimation. (see parameters below).
- Nov 2024: FlexSIM is now compatible with GPU computation (with Matlab Parrallel Toolbox). Set params.GPU = 1 and it's done!
- March 2024: New parameters to ease the handling of temporal stacks.
- Feb 2024: Integration of a pure Matlab version of the VMLMB optimization method in GlobalBioIm which is used in FlexSIM. As such, no need anymore to compile mex files.
Getting started
Download or clone this repository and run the script InstallFlexSIM.m which will download the required GlobalBioIm library v1.2 (or more recent releases) and make neceesary changes in your Matlab path.
Repository content
The repository is organized as follows.
- File InstallFlexSIM.m: FlexSIM installation script
- File FlexSIM.m: Main function of FlexSIM, the one that needs to be run with parameters as input (see provided examples)
- Folder src: Matlab source files of FlexSIM
- Folder Examples: Scripts to download and reconstruct the 20 open datasets described in Table S1 of [1] (see Examples below).
FlexSIM parameters
General parameters
| Parameter | Description |
|---|---|
| DataPath | Path to the SIM raw stack. |
| pathToFlexSIM | Path to FlexSIM root folder. |
| displ | From 0 to 2 with increasing number of display. |
| verbose | From 0 to 2 with increasing text displays. |
| sav | Saving options (0: no saving / 1: save reconstruction / 2: save patterns and reconstruction). |
| dataType | Data type for saving: 'uint8', 'uint16', 'uint32', 'single' (default) |
| GPU | Boolean on whether to use GPU or not |
| parallelProcess | Boolean on whether to use parallel computing. Requires the parallel computing toolbox. |
Physical parameters and pre-processing
| Parameter | Description |
|---|---|
| ---- Reconstruct on ROI / specific frames | |
| SzRoi | Size (px) of the ROI of the data considered for reconstruction |
| posRoi | Position of the top-left corner of the ROI. |
| frameRange | To treat only a subset of temporal frames of the stack (e.g., [1;5]). |
| ---- Patterns | |
| StackOrder | Stack order: ap, pa, apw, paw, wap, wpa, axp, pxa (with a=angles, p=phases, w=widefield). The last two correspond to 9x9 montage with angles (resp phases) in row axp (resp pxa). |
| nbOr | Number of orientations. |
| nbPh | Number of phases. |
| ---- OTF Approximation | |
| lamb | Emission wavelength (nm). |
| res | Pixel size (nm). |
| Na | Objective numerical aperture. |
| damp | damping parameter in [0,1] (1= no damping) to attenuate middle freq in the approx OTF. |
| ---- Background estimation | |
| SzRoiBack | Size (px) of the ROI for background estimation (position automatically detected to minimize the intensity within the ROI). |
Patterns estimation
| Parameter | Description |
|---|---|
| SzRoiPatt | Size (px) of the ROI for pattern estimation. |
| posRoiPatt | Position of the top-left corner of the ROI for pattern estimation. (if empty automatically detected to maximize the intensity within the ROI) |
| maskWF | Radius (as a factor of the cutoff freq.) of the disk used to mask central Fourier frequencies. |
| ringRegionSearch | Lower and upper limits of Fourier ring region to search peaks (given as factor of the cutoff freq.). |
| eqPh | Boolean, if true equally-spaced phases are assumed. |
| eqOrr | Boolean, if true equally-spaced orientation are assumed (constraint imposed in a soft way). |
| estiPattLowFreq | Boolean, if true, estimate the low-freq. component of the patterns. |
| doRefinement | If false, do not performs the refinement step |
| pattAmp | Amplitude a of the pattern (to be adjusted manually). |
| framePattEsti | To use only a subset of frames for estimating a common pattern to all frames (empty to use all frames). Only used when cstTimePatt =1. |
| cstTimePatt | If true, a single set of patterns is estimated and used for all frames. |
| rollMed | Size of the rolling median interval used to adjust patterns parameters accros frames. Can only be used with cstTimePatt=0. |
Image reconstruction
| Parameter | Description |
|---|---|
| ---- OTF Attenuation | |
| OTFAttStr | Strength of the OTF attenuation (in [0,1]). If 0 no OTF attenuation. |
| OTFAttwdth | Width of the OTF attenuation (>0). If 0 no OTF attenuation. |
| ---- Dark-Sectioning (from this paper) | |
| DarkSec | If >0 activate the Dark-sectioning (1 for middle and 2 for strong out-of-focus). |
| DarkSecThres | Threshold used to define a mask on the image that contains the background. |
| ---- Cost function | |
| apodize | Boolean on whether to use apodization on boundaries. |
| sepOrr | Boolean on whether to treat each orientation separately. |
| padSz | Padding size (px) used in the forward operator. |
| mu | Regularization parameter. |
| regType | Regularizer: 1 - 1st-order Tikhonov, 2 - Total Variation, 3 - Good roughness. |
| ---- Optimization | |
| maxIt | Maximum number of iterations (stopping criteria). |
| stepTol | elative error tolerance between two iterates (stopping criteria). |
Examples
This folder contains scripts to download and reconstruct 25 open 2D-SIM datasets. Each script is made of the following two steps
- download raw data and set FlexSIM parameters
- run FlexSIM.m function
The 25 open 2D-SIM datasets are sourced from 8 publications including FairSIM [1], OpenSIM [2], HiFi-SIM [3], ML-SIM [4], JSFR-SIM [5], Direct-SIM [6], PCA-SIM [7], B-SIM [8], and BF-SIM [9]. This corresponds to a collection of SIM data acquired with a diversity of SIM systems and configurations. Each subfolder of the Example folder corresponds to one dataset with the following naming convention
- Reference_SIM-type_Bio-structure
[1] FairSIM, M. Müller, V. Mönkemöller, S. Hennig, W. Hübner, and T. Huser, Open-source image reconstruction of super-resolution structured illumination microscopy data in ImageJ, Nat. Commun., vol. 7, no. 1, no. 1, Mar. 2016.
[2] OpenSIM, A. Lal, C. Shan, and P. Xi, Structured Illumination Microscopy Image Reconstruction Algorithm, IEEE Journal of Selected Topics in Quantum Electronics, vol. 22, no. 4, Jul. 2016.
[3] HiFi-SIM, G. Wen et al., High-fidelity structured illumination microscopy by point-spread-function engineering, Light Sci Appl, vol. 10, no. 1, no. 1, Apr. 2021.
[4] ML-SIM, C. N. Christensen, E. N. Ward, P. Lio, and C. F. Kaminski, ML-SIM: A deep neural network for reconstruction of structured illumination microscopy images, Biomed. Opt. Express, vol. 12, no. 5, May 2021.
[5] JSFR-SIM, Z. Wang et al., High-speed image reconstruction for optically sectioned, super-resolution structured illumination microscopy, Advanced Photonics, vol. 4, no. 2, Mar. 2022.
[6] PCA-SIM, J. Qian, et al., Structured illumination microscopy based on principal component analysis, eLight, vol. 3, no. 1, Feb. 2023.
[7] Direct-SIM, G. Wen et al., Spectrum-optimized direct image reconstruction of super-resolution structured illumination microscopy, PhotoniX, vol. 4, no. 1, June 2023.
[8] B-SIM, A. Saurabh, et al., Approaching maximum resolution in structured illumination microscopy via accurate noise modeling, npj Imaging, vol. 3, no. 5, Jan. 2025.
[9] BF-SIM, Y. Mo et al., Quantitative structured illumination microscopy via a physical model-based background filtering algorithm reveals actin dynamics, Nature Communications, 2023.
Conditions of use
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/ .
Whenever you present or publish results that are based on this repository, please cite:
E. Soubies, A. Nogueron, F. Pelletier, T. Mangeat, C. Leterrier, M. Unser, and D. Sage. Surpassing Light Inhomogeneities in Structured-Illumination Microscopy with FlexSIM. Journal of Microscopy, 2024.