Industrial CT

Industrial Computed 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
🥇 Learned Primal-Dual 0.831 36.42 0.947 ✓ Certified Adler & Oktem, IEEE TMI 2018
🥈 FBPConvNet 0.816 35.81 0.939 ✓ Certified Jin et al., IEEE TIP 2017
🥉 PnP-ADMM 0.740 32.64 0.891 ✓ Certified Venkatakrishnan et al., 2013
4 FDK 0.712 30.16 0.919 ✓ Certified Feldkamp et al., JOSA A 1984

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
🥇 FBPConvNet + gradient 0.743
0.795
33.58 dB / 0.958
0.745
30.51 dB / 0.924
0.689
28.17 dB / 0.884
✓ Certified Jin et al., IEEE TIP 2017
🥈 FDK + gradient 0.697
0.731
28.48 dB / 0.891
0.686
27.53 dB / 0.871
0.674
26.95 dB / 0.857
✓ Certified Feldkamp et al., JOSA A 1984
🥉 PnP-ADMM + gradient 0.692
0.748
30.22 dB / 0.920
0.687
26.78 dB / 0.853
0.641
25.81 dB / 0.827
✓ Certified Venkatakrishnan et al., 2013
4 Learned Primal-Dual + gradient 0.685
0.803
33.86 dB / 0.960
0.671
25.92 dB / 0.830
0.581
22.5 dB / 0.711
✓ Certified Adler & Oktem, IEEE TMI 2018

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 3 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 Learned Primal-Dual + gradient 0.803 33.86 0.96
2 FBPConvNet + gradient 0.795 33.58 0.958
3 PnP-ADMM + gradient 0.748 30.22 0.92
4 FDK + gradient 0.731 28.48 0.891
Spec Ranges (4 parameters)
Parameter Min Max Unit
beam_hardening_order -0.5 1.0
scatter_fraction -0.15 0.3
source_blur -3.0 6.0 pixels
detector_efficiency 0.7 1.15
View Details →
Dev 3 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 FBPConvNet + gradient 0.745 30.51 0.924
2 PnP-ADMM + gradient 0.687 26.78 0.853
3 FDK + gradient 0.686 27.53 0.871
4 Learned Primal-Dual + gradient 0.671 25.92 0.83
Spec Ranges (4 parameters)
Parameter Min Max Unit
beam_hardening_order -0.6 0.9
scatter_fraction -0.18 0.27
source_blur -3.6 5.4 pixels
detector_efficiency 0.73 1.18
Submit Reconstruction View Details →
Hidden 3 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 FBPConvNet + gradient 0.689 28.17 0.884
2 FDK + gradient 0.674 26.95 0.857
3 PnP-ADMM + gradient 0.641 25.81 0.827
4 Learned Primal-Dual + gradient 0.581 22.5 0.711
Spec Ranges (4 parameters)
Parameter Min Max Unit
beam_hardening_order -0.35 1.15
scatter_fraction -0.105 0.345
source_blur -2.1 6.9 pixels
detector_efficiency 0.655 1.105
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(θ) → B(poly) → S(scatter) → D(g, η)

R Radon Transform (θ)
B Beam Hardening (Polychromatic) (poly)
S Scatter Contribution (scatter)
D Detector + Noise (g, η)

Mismatch Parameters

Symbol Parameter Description Nominal Perturbed
β_bh beam_hardening_order Beam hardening polynomial coefficient 0 0.5
f_s scatter_fraction Scatter fraction 0 0.15
σ_src source_blur Source focal spot blur (pixels) 0 3.0
η_det detector_efficiency Detector efficiency variation 1.0 0.85

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.