Recent Advances in Structured Illumination Microscope Super-Resolution Image Reconstruction

Yujun Tang; Linbo Wang; Gang Wen; Hui Li

doi:10.3788/LOP202259.0617009

Journals >Laser & Optoelectronics Progress >Volume 59 >Issue 6 >Page 0617009 > Article

Laser & Optoelectronics Progress
Vol. 59, Issue 6, 0617009 (2022)

Recent Advances in Structured Illumination Microscope Super-Resolution Image Reconstruction

Yujun Tang^1、2, Linbo Wang², Gang Wen², and Hui Li^1、2、*

Author Affiliations

¹School of Biomedical Engineering, University of Science and Technology of China, Suzhou , Jiangsu 215163, China

²Jiangsu Key Laboratory of Medical Optics, Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences, Suzhou , Jiangsu 215163, China

show less

DOI: 10.3788/LOP202259.0617009 Cite this Article Set citation alerts

Yujun Tang, Linbo Wang, Gang Wen, Hui Li. Recent Advances in Structured Illumination Microscope Super-Resolution Image Reconstruction[J]. Laser & Optoelectronics Progress, 2022, 59(6): 0617009 Copy Citation Text

show less

Fig. 1. Principle of two-dimensional structured illumination microscopy

Download full size

Principle of iterative autocorrelation algorithm and comparison of reconstruction results with and without accurate parameter estimation. (a) Original image; (b)(c) principle of iterative autocorrelation algorithm; (d) reconstruction result under wrong spatial frequency (error is about 7.69%); (e) reconstruction result under wrong phase (error is about 8.67%); (f) reconstruction result with accurate parameters

Fig. 2. Principle of iterative autocorrelation algorithm and comparison of reconstruction results with and without accurate parameter estimation. (a) Original image; (b)(c) principle of iterative autocorrelation algorithm; (d) reconstruction result under wrong spatial frequency (error is about 7.69%); (e) reconstruction result under wrong phase (error is about 8.67%); (f) reconstruction result with accurate parameters

Download full size

Fig. 3. Quantitative characterization of thefidelity of HiFi-SIM reconstruction^[57]

Download full size

Fig. 4. Comparison of sidelobe artifacts in reconstructed images of Wide-field, fairSIM, TV-SIM, Hessian-SIM, and HiFi-SIM^[57]

Download full size

Algorithm	Download link	Application（benefits）	Running platform
SIMcheck^［64］	https：//github.com/MicronOxford/SIMcheck	1）Assessing the resolution and image quality；2）identification of sources of errors and artefacts in reconstructed images	FiJi /ImageJ
FairSIM^［48］	https：//github.com/fairsim	SIM super-resolution reconstruction for all sinusoidal illumination modes	FiJi /ImageJ
SIMToolbox^［65］	http：//mmtg.fel.cvut.cz/SIMToolbox	1）Optical sectioning，classical SIM algorithm；2）support MAP-SIM algorithm^［50］	MATLAB
OpenSIM^［23］	https：//github.com/LanMai/OpenSIM	Classical 2D-SIM reconstruction algorithm	MATLAB
Hessian-SIM^［51］	https：//www.nature.com/articles/nbt.4115#Sec21	1）Effective suppression of reconstruction artifacts at low signal-to-noise ratios；2）long-time dynamic imaging	MATLAB
Sparse-SIM^［46］	https：//github.com/WeisongZhao/Sparse-SIM	1）Effective suppression of reconstruction artifacts at low signal-to-noise ratios；2）high spatial resolution（~60 nm）	MATLAB
HiFi-SIM^［57］	https：//doi.org/10.1038/s41377-021-00513-w	1）Effective reduction of reconstruction artifacts；2）effective solution to the problems caused by PSF mismatch	MATLAB
True-Wiener SIM^［58］	https：//github.com/qnano/simnoise	1）Reduction of user adjustable parameters；2）high contrast imaging	MATLAB
Flat-noise SIM^［58］	https：//github.com/qnano/simnoise	1）No user adjustable parameters；2）suppression of structural noise artifacts	MATLAB
Notch-fitered SIM^［58］	https：//github.com/qnano/simnoise	1）No user adjustable parameters；2）higher image contrast than flat-noise SIM	MATLAB

Table 1. Open-source SIM reconstruction algorithms

Algorithm	Network	Download link	Advantage	Biological sample
scU-Net^［78］	U-Net	https：//github.com/drbeiliu/DeepLearning	High quality image reconstruction in low signal-to-noise conditions	Microtubules；adhesions；mitochondria；F-actin
DFCAN/DFGAN^［79］	GAN	https：//github.com/qc17-THU/DL-SR	High quality image reconstruction in low signal-to-noise conditions	Clathrin-coated pits； endoplasmic reticulum；microtubules；F-actin
RED-fairSIM^［80］	RED-Net	—	1）High quality image reconstruction in low signal-to-noise conditions；2）no image pre-processing required；3）low training costs	U2OS cells；tubulin cytoskeleton
ML-SIM^［82］	RCAN	http：//ML-SIM.github.io	1）High quality image reconstruction in low signal-to-noise conditions；2）model is highly generalized	Endoplasmic reticulum；cell membrane
U-Net-SIM3^［78］	U-Net	https：//github.com/drbeiliu/DeepLearning	Fewer raw images（five-fold reduction）	Microtubules；adhesions；mitochondria；F-actin
Ref.［83］	cycleGAN	—	Fewer raw images（three original images）	—
caGAN^［84］	GAN	https：//github.com/qc17-THU/DL-SR	1）Halving the number of raw images（axial）；2）high quality image reconstruction in low signal-to-noise conditions	Clathrin-coated pits； endoplasmic reticulum；microtubules；F-actin
Ref.［76］	GAN	—	1）Large imaging field of view； 2）wide field images for super resolution	Gene-edited SUM159；drosophila embryos

Table 2. Deep learning based SIM reconstruction algorithm

Yujun Tang, Linbo Wang, Gang Wen, Hui Li. Recent Advances in Structured Illumination Microscope Super-Resolution Image Reconstruction[J]. Laser & Optoelectronics Progress, 2022, 59(6): 0617009

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information