Small-Angle X-ray Scattering (SAXS)

Small-Angle X-ray Scattering (SAXS)

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
🥇 ScatterFormer 0.771 33.5 0.925 ✓ Certified Scattering transformer, 2024
🥈 ScatterNet 0.723 31.5 0.895 ✓ Certified Franke et al., 2018
🥉 McSAS 0.603 27.5 0.790 ✓ Certified Bressler et al., 2015
4 PyFAI-Integrate 0.485 24.0 0.670 ✓ Certified Ashiotis et al., 2015

Dataset: PWM Benchmark (4 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
🥇 ScatterFormer + gradient 0.694
0.765
31.7 dB / 0.939
0.680
26.76 dB / 0.852
0.637
25.45 dB / 0.816
✓ Certified Scattering analysis transformer, 2024
🥈 ScatterNet + gradient 0.646
0.757
30.41 dB / 0.923
0.621
23.83 dB / 0.762
0.560
22.36 dB / 0.705
✓ Certified Franke et al., Biophys. J. 2018
🥉 McSAS + gradient 0.627
0.642
24.53 dB / 0.787
0.616
23.85 dB / 0.763
0.624
24.6 dB / 0.789
✓ Certified Bressler et al., J. Appl. Cryst. 2015
4 PyFAI-Integrate + gradient 0.541
0.565
21.73 dB / 0.678
0.534
21.28 dB / 0.658
0.523
20.32 dB / 0.614
✓ Certified Ashiotis et al., J. Appl. Cryst. 2015

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 ScatterFormer + gradient 0.765 31.7 0.939
2 ScatterNet + gradient 0.757 30.41 0.923
3 McSAS + gradient 0.642 24.53 0.787
4 PyFAI-Integrate + gradient 0.565 21.73 0.678
Spec Ranges (4 parameters)
Parameter Min Max Unit
sample_detector_distance 998.0 1004.0 mm
beam_center_x -0.4 0.8 px
beam_center_y -0.4 0.8 px
wavelength_error -0.0002 0.0004 nm
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 ScatterFormer + gradient 0.680 26.76 0.852
2 ScatterNet + gradient 0.621 23.83 0.762
3 McSAS + gradient 0.616 23.85 0.763
4 PyFAI-Integrate + gradient 0.534 21.28 0.658
Spec Ranges (4 parameters)
Parameter Min Max Unit
sample_detector_distance 997.6 1003.6 mm
beam_center_x -0.48 0.72 px
beam_center_y -0.48 0.72 px
wavelength_error -0.00024 0.00036 nm
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 ScatterFormer + gradient 0.637 25.45 0.816
2 McSAS + gradient 0.624 24.6 0.789
3 ScatterNet + gradient 0.560 22.36 0.705
4 PyFAI-Integrate + gradient 0.523 20.32 0.614
Spec Ranges (4 parameters)
Parameter Min Max Unit
sample_detector_distance 998.6 1004.6 mm
beam_center_x -0.28 0.92 px
beam_center_y -0.28 0.92 px
wavelength_error -0.00014 0.00046 nm
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

R → D

R Rotation
D Detector

Mismatch Parameters

Symbol Parameter Description Nominal Perturbed
s_d sample_detector_distance Sample-detector distance (mm) 1000.0 1002.0
b_c beam_center_x Beam center x (px) 0.0 0.4
b_c beam_center_y Beam center y (px) 0.0 0.4
w_e wavelength_error Wavelength error (nm) 0.0 0.0002

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.