Secondary Ion Mass Spectrometry (SIMS) Imaging

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
🥇	U-Net-Spectra U-Net-Spectra Spectral U-Net variant 33.16 dB SSIM 0.954 Checkpoint unavailable	0.780	33.16	0.954	✓ Certified	Spectral U-Net variant
🥈	ScoreSpectra ScoreSpectra Wei et al., 2025 32.48 dB SSIM 0.948 Checkpoint unavailable	0.765	32.48	0.948	✓ Certified	Wei et al., 2025
🥉	Cascade-UNet Cascade-UNet Physics-informed UNet, 2025 33.0 dB SSIM 0.922 Checkpoint unavailable	0.761	33.0	0.922	✓ Certified	Physics-informed UNet, 2025
4	PINN-Spectra PINN-Spectra Physics-informed neural network 31.76 dB SSIM 0.940 Checkpoint unavailable	0.749	31.76	0.940	✓ Certified	Physics-informed neural network
5	SpectraFormer SpectraFormer Spectroscopy transformer, 2024 30.76 dB SSIM 0.928 Checkpoint unavailable	0.727	30.76	0.928	✓ Certified	Spectroscopy transformer, 2024
6	DiffusionSpectra DiffusionSpectra Zhang et al., 2024 30.68 dB SSIM 0.927 Checkpoint unavailable	0.725	30.68	0.927	✓ Certified	Zhang et al., 2024
7	CDAE CDAE Zhang et al., Sensors 2024 31.5 dB SSIM 0.895 Checkpoint unavailable	0.723	31.5	0.895	✓ Certified	Zhang et al., Sensors 2024
8	PnP-DnCNN PnP-DnCNN Zhang et al., 2017 27.9 dB SSIM 0.800 Try in SpecLab →	0.615	27.9	0.800	✓ Certified	Zhang et al., 2017
9	Baseline Correction Baseline Correction Polynomial fitting baseline 25.01 dB SSIM 0.803 Try in SpecLab →	0.568	25.01	0.803	✓ Certified	Polynomial fitting baseline
10	SVD SVD Singular Value Decomposition 23.94 dB SSIM 0.766 Try in SpecLab →	0.532	23.94	0.766	✓ Certified	Singular Value Decomposition
11	SG-ALS SG-ALS Savitzky-Golay + ALS baseline 24.3 dB SSIM 0.670 Try in SpecLab →	0.490	24.3	0.670	✓ Certified	Savitzky-Golay + ALS baseline

Dataset: PWM Benchmark (11 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
🥇	U-Net-Spectra + gradient U-Net-Spectra + gradient Spectral U-Net variant Score 0.660 Correct & Reconstruct →	0.660	0.760 31.21 dB / 0.934	0.649 25.01 dB / 0.803	0.572 22.22 dB / 0.699	✓ Certified	Spectral U-Net variant
🥈	PINN-Spectra + gradient PINN-Spectra + gradient Physics-informed neural network Score 0.648 Correct & Reconstruct →	0.648	0.734 29.46 dB / 0.908	0.643 25.44 dB / 0.816	0.566 22.15 dB / 0.696	✓ Certified	Physics-informed neural network
🥉	CDAE + gradient CDAE + gradient Zhang et al., Sensors 2024 Score 0.640 Correct & Reconstruct →	0.640	0.756 30.33 dB / 0.922	0.614 23.94 dB / 0.766	0.549 21.9 dB / 0.686	✓ Certified	Zhang et al., Sensors 2024
4	ScoreSpectra + gradient ScoreSpectra + gradient Wei et al., 2025 Score 0.634 Correct & Reconstruct →	0.634	0.772 31.37 dB / 0.935	0.605 23.89 dB / 0.765	0.525 20.41 dB / 0.618	✓ Certified	Wei et al., 2025
5	PnP-DnCNN + gradient PnP-DnCNN + gradient Zhang et al., 2017 Score 0.633 Correct & Reconstruct →	0.633	0.695 26.88 dB / 0.855	0.634 24.78 dB / 0.795	0.571 22.71 dB / 0.720	✓ Certified	Zhang et al., 2017
6	Cascade-UNet + gradient Cascade-UNet + gradient Physics-informed UNet, 2025 Score 0.629 Correct & Reconstruct →	0.629	0.752 30.2 dB / 0.920	0.599 23.17 dB / 0.738	0.535 21.52 dB / 0.669	✓ Certified	Physics-informed UNet, 2025
7	DiffusionSpectra + gradient DiffusionSpectra + gradient Zhang et al., 2024 Score 0.629 Correct & Reconstruct →	0.629	0.716 28.31 dB / 0.887	0.611 23.5 dB / 0.750	0.560 22.45 dB / 0.709	✓ Certified	Zhang et al., 2024
8	SpectraFormer + gradient SpectraFormer + gradient Spectroscopy transformer, 2024 Score 0.591 Correct & Reconstruct →	0.591	0.720 28.78 dB / 0.896	0.574 22.16 dB / 0.697	0.478 19.54 dB / 0.576	✓ Certified	Spectroscopy transformer, 2024
9	Baseline Correction + gradient Baseline Correction + gradient Polynomial fitting baseline Score 0.553 Correct & Reconstruct →	0.553	0.587 22.42 dB / 0.708	0.547 21.67 dB / 0.676	0.526 20.37 dB / 0.616	✓ Certified	Polynomial fitting baseline
10	SG-ALS + gradient SG-ALS + gradient Savitzky-Golay + ALS baseline Score 0.552 Correct & Reconstruct →	0.552	0.577 22.17 dB / 0.697	0.544 21.15 dB / 0.652	0.534 21.33 dB / 0.661	✓ Certified	Savitzky-Golay + ALS baseline
11	SVD + gradient SVD + gradient Singular Value Decomposition Score 0.520 Correct & Reconstruct →	0.520	0.565 21.73 dB / 0.678	0.502 20.24 dB / 0.610	0.492 19.59 dB / 0.579	✓ Certified	Singular Value Decomposition

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	ScoreSpectra + gradient	0.772	31.37	0.935
2	U-Net-Spectra + gradient	0.760	31.21	0.934
3	CDAE + gradient	0.756	30.33	0.922
4	Cascade-UNet + gradient	0.752	30.2	0.92
5	PINN-Spectra + gradient	0.734	29.46	0.908
6	SpectraFormer + gradient	0.720	28.78	0.896
7	DiffusionSpectra + gradient	0.716	28.31	0.887
8	PnP-DnCNN + gradient	0.695	26.88	0.855
9	Baseline Correction + gradient	0.587	22.42	0.708
10	SG-ALS + gradient	0.577	22.17	0.697
11	SVD + gradient	0.565	21.73	0.678

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
mass_calibration_drift	-1.0	2.0	ppm
matrix_effect_(sputter_yield)	-10.0	20.0	-
crater_edge_effect	-2.0	4.0	-
charging_(insulating_samples)	-40.0	80.0	V

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	U-Net-Spectra + gradient	0.649	25.01	0.803
2	PINN-Spectra + gradient	0.643	25.44	0.816
3	PnP-DnCNN + gradient	0.634	24.78	0.795
4	CDAE + gradient	0.614	23.94	0.766
5	DiffusionSpectra + gradient	0.611	23.5	0.75
6	ScoreSpectra + gradient	0.605	23.89	0.765
7	Cascade-UNet + gradient	0.599	23.17	0.738
8	SpectraFormer + gradient	0.574	22.16	0.697
9	Baseline Correction + gradient	0.547	21.67	0.676
10	SG-ALS + gradient	0.544	21.15	0.652
11	SVD + gradient	0.502	20.24	0.61

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
mass_calibration_drift	-1.2	1.8	ppm
matrix_effect_(sputter_yield)	-12.0	18.0	-
crater_edge_effect	-2.4	3.6	-
charging_(insulating_samples)	-48.0	72.0	V

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	U-Net-Spectra + gradient	0.572	22.22	0.699
2	PnP-DnCNN + gradient	0.571	22.71	0.72
3	PINN-Spectra + gradient	0.566	22.15	0.696
4	DiffusionSpectra + gradient	0.560	22.45	0.709
5	CDAE + gradient	0.549	21.9	0.686
6	Cascade-UNet + gradient	0.535	21.52	0.669
7	SG-ALS + gradient	0.534	21.33	0.661
8	Baseline Correction + gradient	0.526	20.37	0.616
9	ScoreSpectra + gradient	0.525	20.41	0.618
10	SVD + gradient	0.492	19.59	0.579
11	SpectraFormer + gradient	0.478	19.54	0.576

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
mass_calibration_drift	-0.7	2.3	ppm
matrix_effect_(sputter_yield)	-7.0	23.0	-
crater_edge_effect	-1.4	4.6	-
charging_(insulating_samples)	-28.0	92.0	V

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

S → D

S Sampling

D Detector

Mismatch Parameters

Symbol	Parameter	Description	Perturbed
m_c	mass_calibration_drift	Mass calibration drift (ppm)	1.0
m_e	matrix_effect_(sputter_yield)	Matrix effect (sputter yield) (-)	10.0
c_e	crater_edge_effect	Crater edge effect (-)	2.0
c_(	charging_(insulating_samples)	Charging (insulating samples) (V)	40.0

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.