US/MRI Fusion

US/MRI Fusion

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
🥇 TransMorph 0.799 34.11 0.962 ✓ Certified Chen et al., Med. Image Anal. 2022
🥈 VoxelMorph 0.753 31.93 0.942 ✓ Certified Balakrishnan et al., IEEE TMI 2019
🥉 Demons 0.573 25.16 0.807 ✓ Certified Thirion, Med. Image Anal. 1998
4 B-spline FFD 0.571 25.08 0.805 ✓ Certified Rueckert et al., IEEE TMI 1999

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
🥇 TransMorph + gradient 0.691
0.774
32.22 dB / 0.945
0.691
27.51 dB / 0.870
0.608
24.02 dB / 0.769
✓ Certified Chen et al., Med. Image Anal. 2022
🥈 B-spline FFD + gradient 0.595
0.629
23.89 dB / 0.765
0.600
23.55 dB / 0.752
0.557
21.75 dB / 0.679
✓ Certified Rueckert et al., IEEE TMI 1999
🥉 VoxelMorph + gradient 0.591
0.735
29.14 dB / 0.903
0.572
22.0 dB / 0.690
0.467
18.64 dB / 0.532
✓ Certified Balakrishnan et al., IEEE TMI 2019
4 Demons + gradient 0.589
0.628
23.68 dB / 0.757
0.586
23.04 dB / 0.733
0.554
22.14 dB / 0.696
✓ Certified Thirion, Med. Image Anal. 1998

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 TransMorph + gradient 0.774 32.22 0.945
2 VoxelMorph + gradient 0.735 29.14 0.903
3 B-spline FFD + gradient 0.629 23.89 0.765
4 Demons + gradient 0.628 23.68 0.757
Spec Ranges (3 parameters)
Parameter Min Max Unit
registration_error_(deformable) -2.0 4.0 mm
probe_pressure_deformation -3.0 6.0 mm
mr_distortion -1.0 2.0 mm
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 TransMorph + gradient 0.691 27.51 0.87
2 B-spline FFD + gradient 0.600 23.55 0.752
3 Demons + gradient 0.586 23.04 0.733
4 VoxelMorph + gradient 0.572 22.0 0.69
Spec Ranges (3 parameters)
Parameter Min Max Unit
registration_error_(deformable) -2.4 3.6 mm
probe_pressure_deformation -3.6 5.4 mm
mr_distortion -1.2 1.8 mm
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 TransMorph + gradient 0.608 24.02 0.769
2 B-spline FFD + gradient 0.557 21.75 0.679
3 Demons + gradient 0.554 22.14 0.696
4 VoxelMorph + gradient 0.467 18.64 0.532
Spec Ranges (3 parameters)
Parameter Min Max Unit
registration_error_(deformable) -1.4 4.6 mm
probe_pressure_deformation -2.1 6.9 mm
mr_distortion -0.7 2.3 mm
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

(P → D) + (M → F → S → D) → ⊕

P Propagation
D Detector (US)
M Modulation
F Fourier
S Sampling
D Detector (MR)
Fusion Fusion

Mismatch Parameters

Symbol Parameter Description Nominal Perturbed
r_e registration_error_(deformable) Registration error (deformable) (mm) 0.0 2.0
p_p probe_pressure_deformation Probe pressure deformation (mm) 0.0 3.0
m_d mr_distortion MR distortion (mm) 0.0 1.0

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.