Second Harmonic Generation (SHG) Microscopy

Join the Competition Contribute to Benchmark

Standard reconstruction benchmark — forward model perfectly known, no calibration needed. Score = 0.5 × clip((PSNR−15)/30, 0, 1) + 0.5 × SSIM

#	Method	Score	PSNR (dB)	SSIM	Source
🥇	ScoreMicro ScoreMicro Wei et al., ECCV 2025 38.48 dB SSIM 0.981 Checkpoint unavailable	0.882	38.48	0.981	✓ Certified	Wei et al., ECCV 2025
🥈	DiffDeconv DiffDeconv Huang et al., NeurIPS 2024 38.12 dB SSIM 0.979 Checkpoint unavailable	0.875	38.12	0.979	✓ Certified	Huang et al., NeurIPS 2024
🥉	Restormer+ Restormer+ Zamir et al., ICCV 2024 37.65 dB SSIM 0.975 Checkpoint unavailable	0.865	37.65	0.975	✓ Certified	Zamir et al., ICCV 2024
4	DeconvFormer DeconvFormer Chen et al., CVPR 2024 37.25 dB SSIM 0.972 Checkpoint unavailable	0.857	37.25	0.972	✓ Certified	Chen et al., CVPR 2024
5	ResUNet ResUNet DeCelle et al., Nat. Methods 2021 35.85 dB SSIM 0.964 Checkpoint unavailable	0.830	35.85	0.964	✓ Certified	DeCelle et al., Nat. Methods 2021
6	Restormer Restormer Zamir et al., CVPR 2022 35.8 dB SSIM 0.962 Checkpoint unavailable	0.828	35.8	0.962	✓ Certified	Zamir et al., CVPR 2022
7	U-Net U-Net Ronneberger et al., MICCAI 2015 35.15 dB SSIM 0.956 Checkpoint unavailable	0.814	35.15	0.956	✓ Certified	Ronneberger et al., MICCAI 2015
8	CARE CARE Weigert et al., Nat. Methods 2018 34.5 dB SSIM 0.948 Checkpoint unavailable	0.799	34.5	0.948	✓ Certified	Weigert et al., Nat. Methods 2018
9	PnP-DnCNN PnP-DnCNN Zhang et al., IEEE TIP 2017 31.2 dB SSIM 0.890 Try in SpecLab →	0.715	31.2	0.890	✓ Certified	Zhang et al., IEEE TIP 2017
10	PnP-FISTA PnP-FISTA Bai et al., 2020 30.42 dB SSIM 0.872 Try in SpecLab →	0.693	30.42	0.872	✓ Certified	Bai et al., 2020
11	TV-Deconvolution TV-Deconvolution TV-regularized deconvolution 29.5 dB SSIM 0.845 Try in SpecLab →	0.664	29.5	0.845	✓ Certified	TV-regularized deconvolution
12	Wiener Filter Wiener Filter Analytical baseline 28.35 dB SSIM 0.805 Try in SpecLab →	0.625	28.35	0.805	✓ Certified	Analytical baseline
13	Richardson-Lucy Richardson-Lucy Richardson 1972 / Lucy 1974 27.1 dB SSIM 0.770 Try in SpecLab →	0.587	27.1	0.770	✓ Certified	Richardson 1972 / Lucy 1974

Dataset: PWM Benchmark (13 algorithms)

Blind Reconstruction Challenge — forward model has unknown mismatch, must calibrate from data. Score = 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)

#	Method	Overall Score	Public PSNR / SSIM	Dev PSNR / SSIM	Hidden PSNR / SSIM	Trust	Source
🥇	Restormer+ + gradient Restormer+ + gradient Zamir et al., ICCV 2024 Score 0.787 Correct & Reconstruct →	0.787	0.816 34.74 dB / 0.966	0.784 32.89 dB / 0.952	0.762 32.31 dB / 0.946	✓ Certified	Zamir et al., ICCV 2024
🥈	ScoreMicro + gradient ScoreMicro + gradient Wei et al., ECCV 2025 Score 0.764 Correct & Reconstruct →	0.764	0.829 36.32 dB / 0.975	0.753 31.3 dB / 0.935	0.711 29.02 dB / 0.901	✓ Certified	Wei et al., ECCV 2025
🥉	DeconvFormer + gradient DeconvFormer + gradient Chen et al., CVPR 2024 Score 0.761 Correct & Reconstruct →	0.761	0.812 34.6 dB / 0.965	0.770 32.23 dB / 0.945	0.700 27.95 dB / 0.880	✓ Certified	Chen et al., CVPR 2024
4	ResUNet + gradient ResUNet + gradient DeCelle et al., Nat. Methods 2021 Score 0.731 Correct & Reconstruct →	0.731	0.795 33.27 dB / 0.955	0.727 29.06 dB / 0.901	0.670 27.46 dB / 0.869	✓ Certified	DeCelle et al., Nat. Methods 2021
5	Restormer + gradient Restormer + gradient Zamir et al., CVPR 2022 Score 0.715 Correct & Reconstruct →	0.715	0.792 32.9 dB / 0.952	0.727 28.93 dB / 0.899	0.627 25.27 dB / 0.811	✓ Certified	Zamir et al., CVPR 2022
6	DiffDeconv + gradient DiffDeconv + gradient Huang et al., NeurIPS 2024 Score 0.714 Correct & Reconstruct →	0.714	0.825 36.29 dB / 0.975	0.694 27.97 dB / 0.880	0.622 24.17 dB / 0.775	✓ Certified	Huang et al., NeurIPS 2024
7	PnP-DnCNN + gradient PnP-DnCNN + gradient Zhang et al., IEEE TIP 2017 Score 0.702 Correct & Reconstruct →	0.702	0.750 29.82 dB / 0.914	0.693 27.51 dB / 0.870	0.662 26.34 dB / 0.841	✓ Certified	Zhang et al., IEEE TIP 2017
8	U-Net + gradient U-Net + gradient Ronneberger et al., MICCAI 2015 Score 0.702 Correct & Reconstruct →	0.702	0.785 32.69 dB / 0.950	0.683 27.2 dB / 0.863	0.637 25.36 dB / 0.813	✓ Certified	Ronneberger et al., MICCAI 2015
9	CARE + gradient CARE + gradient Weigert et al., Nat. Methods 2018 Score 0.699 Correct & Reconstruct →	0.699	0.773 31.6 dB / 0.938	0.691 27.07 dB / 0.860	0.632 24.78 dB / 0.795	✓ Certified	Weigert et al., Nat. Methods 2018
10	TV-Deconvolution + gradient TV-Deconvolution + gradient Rudin et al., Phys. A 1992 Score 0.688 Correct & Reconstruct →	0.688	0.718 27.85 dB / 0.878	0.686 27.16 dB / 0.862	0.659 26.12 dB / 0.835	✓ Certified	Rudin et al., Phys. A 1992
11	PnP-FISTA + gradient PnP-FISTA + gradient Bai et al., 2020 Score 0.674 Correct & Reconstruct →	0.674	0.738 29.23 dB / 0.904	0.668 26.45 dB / 0.844	0.617 24.32 dB / 0.780	✓ Certified	Bai et al., 2020
12	Wiener Filter + gradient Wiener Filter + gradient Analytical baseline Score 0.653 Correct & Reconstruct →	0.653	0.664 25.5 dB / 0.818	0.651 25.94 dB / 0.830	0.643 25.45 dB / 0.816	✓ Certified	Analytical baseline
13	Richardson-Lucy + gradient Richardson-Lucy + gradient Richardson, JOSA 1972 / Lucy, AJ 1974 Score 0.616 Correct & Reconstruct →	0.616	0.675 25.9 dB / 0.829	0.629 24.56 dB / 0.788	0.544 21.94 dB / 0.687	✓ Certified	Richardson, JOSA 1972 / Lucy, AJ 1974

Complete score requires all 3 tiers (Public + Dev + Hidden).

Join the competition →

Scoring: 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖) PSNR 40% · SSIM 40% · Consistency 20%

Public 5 scenes

Full-access development tier with all data visible.

What you get & how to use

What you get: Measurements (y), ideal forward operator (H), spec ranges, ground truth (x_true), and true mismatch spec.

How to use: Load HDF5 → compare reconstruction vs x_true → check consistency → iterate.

What to submit: Reconstructed signals (x_hat) and corrected spec as HDF5.

Public Leaderboard

#	Method	Score	PSNR	SSIM
1	ScoreMicro + gradient	0.829	36.32	0.975
2	DiffDeconv + gradient	0.825	36.29	0.975
3	Restormer+ + gradient	0.816	34.74	0.966
4	DeconvFormer + gradient	0.812	34.6	0.965
5	ResUNet + gradient	0.795	33.27	0.955
6	Restormer + gradient	0.792	32.9	0.952
7	U-Net + gradient	0.785	32.69	0.95
8	CARE + gradient	0.773	31.6	0.938
9	PnP-DnCNN + gradient	0.750	29.82	0.914
10	PnP-FISTA + gradient	0.738	29.23	0.904
11	TV-Deconvolution + gradient	0.718	27.85	0.878
12	Richardson-Lucy + gradient	0.675	25.9	0.829
13	Wiener Filter + gradient	0.664	25.5	0.818

Spec Ranges (3 parameters)

Parameter	Min	Max	Unit
phase_matching_error	-1.0	2.0	-
excitation_power_fluctuation	-2.0	4.0	-
collection_na_mismatch	-0.02	0.04	-

View Details →

Dev 5 scenes

Blind evaluation tier — no ground truth available.

What you get & how to use

What you get: Measurements (y), ideal forward operator (H), and spec ranges only.

How to use: Apply your pipeline from the Public tier. Use consistency as self-check.

What to submit: Reconstructed signals and corrected spec. Scored server-side.

Dev Leaderboard

#	Method	Score	PSNR	SSIM
1	Restormer+ + gradient	0.784	32.89	0.952
2	DeconvFormer + gradient	0.770	32.23	0.945
3	ScoreMicro + gradient	0.753	31.3	0.935
4	ResUNet + gradient	0.727	29.06	0.901
5	Restormer + gradient	0.727	28.93	0.899
6	DiffDeconv + gradient	0.694	27.97	0.88
7	PnP-DnCNN + gradient	0.693	27.51	0.87
8	CARE + gradient	0.691	27.07	0.86
9	TV-Deconvolution + gradient	0.686	27.16	0.862
10	U-Net + gradient	0.683	27.2	0.863
11	PnP-FISTA + gradient	0.668	26.45	0.844
12	Wiener Filter + gradient	0.651	25.94	0.83
13	Richardson-Lucy + gradient	0.629	24.56	0.788

Spec Ranges (3 parameters)

Parameter	Min	Max	Unit
phase_matching_error	-1.2	1.8	-
excitation_power_fluctuation	-2.4	3.6	-
collection_na_mismatch	-0.024	0.036	-

Submit Reconstruction View Details →

Hidden 5 scenes

Fully blind server-side evaluation — no data download.

What you get & how to use

What you get: No data downloadable. Algorithm runs server-side on hidden measurements.

How to use: Package algorithm as Docker container / Python script. Submit via link.

What to submit: Containerized algorithm accepting y + H, outputting x_hat + corrected spec.

Hidden Leaderboard

#	Method	Score	PSNR	SSIM
1	Restormer+ + gradient	0.762	32.31	0.946
2	ScoreMicro + gradient	0.711	29.02	0.901
3	DeconvFormer + gradient	0.700	27.95	0.88
4	ResUNet + gradient	0.670	27.46	0.869
5	PnP-DnCNN + gradient	0.662	26.34	0.841
6	TV-Deconvolution + gradient	0.659	26.12	0.835
7	Wiener Filter + gradient	0.643	25.45	0.816
8	U-Net + gradient	0.637	25.36	0.813
9	CARE + gradient	0.632	24.78	0.795
10	Restormer + gradient	0.627	25.27	0.811
11	DiffDeconv + gradient	0.622	24.17	0.775
12	PnP-FISTA + gradient	0.617	24.32	0.78
13	Richardson-Lucy + gradient	0.544	21.94	0.687

Spec Ranges (3 parameters)

Parameter	Min	Max	Unit
phase_matching_error	-0.7	2.3	-
excitation_power_fluctuation	-1.4	4.6	-
collection_na_mismatch	-0.014	0.046	-

Submit Algorithm View Details →

Blind Reconstruction Challenge

Challenge

Given measurements with unknown mismatch and spec ranges (not exact params), reconstruct the original signal. A method must be evaluated on all three tiers for a complete score. Scored on a composite metric: 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖).

Input

Measurements y, ideal forward model H, spec ranges

Output

Reconstructed signal x̂

Spec DAG — Forward Model Pipeline

M → R → D

M Modulation

R Rotation

D Detector

Mismatch Parameters

Symbol	Parameter	Description	Perturbed
p_m	phase_matching_error	Phase matching error (-)	1.0
e_p	excitation_power_fluctuation	Excitation power fluctuation (-)	2.0
c_n	collection_na_mismatch	Collection NA mismatch (-)	0.02

Credits System

40%

Platform Profit Pool

Revenue allocated to benchmark rewards

30%

Winner Share

Top algorithm receives from pool

$100

Min Withdrawal

Minimum payout threshold

Spec Primitives Reference (11 primitives)

P Propagation

Free-space or medium propagation kernel (Fresnel, Rayleigh-Sommerfeld).

M Mask / Modulation

Spatial or spatio-temporal amplitude modulation (coded aperture, SLM pattern).

Π Projection

Geometric projection operator (Radon transform, fan-beam, cone-beam).

F Fourier Sampling

Sampling in the Fourier / k-space domain (MRI, ptychography).

C Convolution

Shift-invariant convolution with a point-spread function (PSF).

Σ Summation / Integration

Summation along a physical dimension (spectral, temporal, angular).

D Detector

Sensor readout with gain g and noise model η (Gaussian, Poisson, mixed).

S Structured Illumination

Patterned illumination (block, Hadamard, random) applied to the scene.

W Wavelength Dispersion

Spectral dispersion element (prism, grating) with shift α and aperture a.

R Rotation / Motion

Sample or gantry rotation (CT, electron tomography).

Λ Wavelength Selection

Spectral filter or monochromator selecting a wavelength band.