X-ray Fluorescence Tomography

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
🥇	DiffusionInstrumentation DiffusionInstrumentation Zhang et al., 2024 33.26 dB SSIM 0.955 Checkpoint unavailable	0.782	33.26	0.955	✓ Certified	Zhang et al., 2024
🥈	Instrument-CNN Instrument-CNN Instrument-specific CNN 32.31 dB SSIM 0.946 Checkpoint unavailable	0.762	32.31	0.946	✓ Certified	Instrument-specific CNN
🥉	CalibFormer CalibFormer Transformer calibration, 2024 32.8 dB SSIM 0.920 Checkpoint unavailable	0.757	32.8	0.920	✓ Certified	Transformer calibration, 2024
4	MassSpecFormer MassSpecFormer Mass spectrometry transformer, 2024 30.98 dB SSIM 0.931 Checkpoint unavailable	0.732	30.98	0.931	✓ Certified	Mass spectrometry transformer, 2024
5	ResNet-Calib ResNet-Calib ResNet for calibration, 2022 31.3 dB SSIM 0.892 Checkpoint unavailable	0.718	31.3	0.892	✓ Certified	ResNet for calibration, 2022
6	ScoreInstrumentation ScoreInstrumentation Wei et al., 2025 29.05 dB SSIM 0.901 Checkpoint unavailable	0.685	29.05	0.901	✓ Certified	Wei et al., 2025
7	PnP-NLM PnP-NLM Non-local means filter 26.92 dB SSIM 0.856 Try in SpecLab →	0.627	26.92	0.856	✓ Certified	Non-local means filter
8	Calibration-Lookup Calibration-Lookup Look-up table calibration 26.73 dB SSIM 0.851 Try in SpecLab →	0.621	26.73	0.851	✓ Certified	Look-up table calibration
9	PnP-BM3D PnP-BM3D Danielyan et al., 2012 27.6 dB SSIM 0.790 Try in SpecLab →	0.605	27.6	0.790	✓ Certified	Danielyan et al., 2012
10	Peak Fitting Peak Fitting Gaussian peak fitting 25.81 dB SSIM 0.827 Try in SpecLab →	0.594	25.81	0.827	✓ Certified	Gaussian peak fitting
11	Deconv Deconv Analytical baseline 24.1 dB SSIM 0.660 Try in SpecLab →	0.482	24.1	0.660	✓ Certified	Analytical 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
🥇	CalibFormer + gradient CalibFormer + gradient Transformer calibration, 2024 Score 0.708 Correct & Reconstruct →	0.708	0.775 31.34 dB / 0.935	0.707 27.74 dB / 0.875	0.643 25.81 dB / 0.827	✓ Certified	Transformer calibration, 2024
🥈	Instrument-CNN + gradient Instrument-CNN + gradient Instrument-specific CNN Score 0.663 Correct & Reconstruct →	0.663	0.741 29.54 dB / 0.910	0.639 25.28 dB / 0.811	0.608 23.84 dB / 0.763	✓ Certified	Instrument-specific CNN
🥉	ResNet-Calib + gradient ResNet-Calib + gradient ResNet for calibration, 2022 Score 0.656 Correct & Reconstruct →	0.656	0.728 29.08 dB / 0.902	0.629 24.42 dB / 0.783	0.612 24.27 dB / 0.778	✓ Certified	ResNet for calibration, 2022
4	MassSpecFormer + gradient MassSpecFormer + gradient Mass spectrometry transformer, 2024 Score 0.656 Correct & Reconstruct →	0.656	0.722 28.75 dB / 0.896	0.657 25.65 dB / 0.822	0.588 23.62 dB / 0.755	✓ Certified	Mass spectrometry transformer, 2024
5	DiffusionInstrumentation + gradient DiffusionInstrumentation + gradient Zhang et al., 2024 Score 0.656 Correct & Reconstruct →	0.656	0.781 31.57 dB / 0.938	0.622 24.11 dB / 0.772	0.564 21.83 dB / 0.683	✓ Certified	Zhang et al., 2024
6	Calibration-Lookup + gradient Calibration-Lookup + gradient Look-up table calibration Score 0.624 Correct & Reconstruct →	0.624	0.625 23.83 dB / 0.762	0.620 23.92 dB / 0.766	0.628 24.78 dB / 0.795	✓ Certified	Look-up table calibration
7	Peak Fitting + gradient Peak Fitting + gradient Gaussian peak fitting Score 0.606 Correct & Reconstruct →	0.606	0.619 23.91 dB / 0.765	0.620 24.17 dB / 0.775	0.578 22.98 dB / 0.730	✓ Certified	Gaussian peak fitting
8	PnP-BM3D + gradient PnP-BM3D + gradient Danielyan et al., 2012 Score 0.599 Correct & Reconstruct →	0.599	0.653 25.17 dB / 0.808	0.594 23.46 dB / 0.749	0.549 21.32 dB / 0.660	✓ Certified	Danielyan et al., 2012
9	Deconv + gradient Deconv + gradient Analytical baseline Score 0.555 Correct & Reconstruct →	0.555	0.599 22.57 dB / 0.714	0.549 21.73 dB / 0.678	0.517 20.9 dB / 0.641	✓ Certified	Analytical baseline
10	ScoreInstrumentation + gradient ScoreInstrumentation + gradient Wei et al., 2025 Score 0.521 Correct & Reconstruct →	0.521	0.688 27.02 dB / 0.859	0.461 18.29 dB / 0.514	0.415 16.7 dB / 0.435	✓ Certified	Wei et al., 2025
11	PnP-NLM + gradient PnP-NLM + gradient Non-local means filter Score 0.520 Correct & Reconstruct →	0.520	0.638 24.53 dB / 0.787	0.508 19.87 dB / 0.592	0.415 16.96 dB / 0.448	✓ Certified	Non-local means filter

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	DiffusionInstrumentation + gradient	0.781	31.57	0.938
2	CalibFormer + gradient	0.775	31.34	0.935
3	Instrument-CNN + gradient	0.741	29.54	0.91
4	ResNet-Calib + gradient	0.728	29.08	0.902
5	MassSpecFormer + gradient	0.722	28.75	0.896
6	ScoreInstrumentation + gradient	0.688	27.02	0.859
7	PnP-BM3D + gradient	0.653	25.17	0.808
8	PnP-NLM + gradient	0.638	24.53	0.787
9	Calibration-Lookup + gradient	0.625	23.83	0.762
10	Peak Fitting + gradient	0.619	23.91	0.765
11	Deconv + gradient	0.599	22.57	0.714

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
self_absorption_correction	-6.0	12.0	-
rotation_axis_offset	-0.6	1.2	px
fluorescence_yield_error	-2.0	4.0	-
dead_time_at_high_count_rate	-2.0	4.0	-

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	CalibFormer + gradient	0.707	27.74	0.875
2	MassSpecFormer + gradient	0.657	25.65	0.822
3	Instrument-CNN + gradient	0.639	25.28	0.811
4	ResNet-Calib + gradient	0.629	24.42	0.783
5	DiffusionInstrumentation + gradient	0.622	24.11	0.772
6	Calibration-Lookup + gradient	0.620	23.92	0.766
7	Peak Fitting + gradient	0.620	24.17	0.775
8	PnP-BM3D + gradient	0.594	23.46	0.749
9	Deconv + gradient	0.549	21.73	0.678
10	PnP-NLM + gradient	0.508	19.87	0.592
11	ScoreInstrumentation + gradient	0.461	18.29	0.514

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
self_absorption_correction	-7.2	10.8	-
rotation_axis_offset	-0.72	1.08	px
fluorescence_yield_error	-2.4	3.6	-
dead_time_at_high_count_rate	-2.4	3.6	-

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	CalibFormer + gradient	0.643	25.81	0.827
2	Calibration-Lookup + gradient	0.628	24.78	0.795
3	ResNet-Calib + gradient	0.612	24.27	0.778
4	Instrument-CNN + gradient	0.608	23.84	0.763
5	MassSpecFormer + gradient	0.588	23.62	0.755
6	Peak Fitting + gradient	0.578	22.98	0.73
7	DiffusionInstrumentation + gradient	0.564	21.83	0.683
8	PnP-BM3D + gradient	0.549	21.32	0.66
9	Deconv + gradient	0.517	20.9	0.641
10	ScoreInstrumentation + gradient	0.415	16.7	0.435
11	PnP-NLM + gradient	0.415	16.96	0.448

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
self_absorption_correction	-4.2	13.8	-
rotation_axis_offset	-0.42	1.38	px
fluorescence_yield_error	-1.4	4.6	-
dead_time_at_high_count_rate	-1.4	4.6	-

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

Π → R → D

Π Projection

R Rotation

D Detector

Mismatch Parameters

Symbol	Parameter	Description	Perturbed
s_c	self_absorption_correction	Self-absorption correction (-)	6.0
r_a	rotation_axis_offset	Rotation axis offset (px)	0.6
f_y	fluorescence_yield_error	Fluorescence yield error (-)	2.0
d_t	dead_time_at_high_count_rate	Dead time at high count rate (-)	2.0

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.