Radio Interferometry (VLBI)

Radio Interferometry (VLBI)

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
🥇 PRIMO 0.722 31.2 0.905 ✓ Certified Medeiros et al., ApJL 2023
🥈 R2D2 0.684 29.8 0.875 ✓ Certified Aghabiglou et al., ApJS 2024
🥉 AIRI 0.573 26.3 0.770 ✓ Certified Terris et al., MNRAS 2022
4 CLEAN 0.425 22.5 0.600 ✓ Certified Hogbom, A&AS 1974

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
🥇 R2D2 + gradient 0.603
0.702
27.79 dB / 0.876
0.580
22.21 dB / 0.699
0.528
20.51 dB / 0.623
✓ Certified Aghabiglou et al., ApJS 2024
🥈 PRIMO + gradient 0.569
0.752
30.1 dB / 0.918
0.519
20.68 dB / 0.631
0.436
17.45 dB / 0.473
✓ Certified Medeiros et al., ApJL 2023
🥉 AIRI + gradient 0.560
0.632
24.52 dB / 0.787
0.564
21.99 dB / 0.690
0.483
19.42 dB / 0.571
✓ Certified Terris et al., MNRAS 2022
4 CLEAN + gradient 0.477
0.510
19.61 dB / 0.580
0.502
19.95 dB / 0.596
0.420
17.12 dB / 0.456
✓ Certified Hogbom, A&AS 1974

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 PRIMO + gradient 0.752 30.1 0.918
2 R2D2 + gradient 0.702 27.79 0.876
3 AIRI + gradient 0.632 24.52 0.787
4 CLEAN + gradient 0.510 19.61 0.58
Spec Ranges (4 parameters)
Parameter Min Max Unit
baseline_error -0.002 0.004 m
phase_calibration -2.0 4.0 deg
amplitude_calibration 0.98 1.04 -
clock_offset -0.2 0.4 ns
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 R2D2 + gradient 0.580 22.21 0.699
2 AIRI + gradient 0.564 21.99 0.69
3 PRIMO + gradient 0.519 20.68 0.631
4 CLEAN + gradient 0.502 19.95 0.596
Spec Ranges (4 parameters)
Parameter Min Max Unit
baseline_error -0.0024 0.0036 m
phase_calibration -2.4 3.6 deg
amplitude_calibration 0.976 1.036 -
clock_offset -0.24 0.36 ns
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 R2D2 + gradient 0.528 20.51 0.623
2 AIRI + gradient 0.483 19.42 0.571
3 PRIMO + gradient 0.436 17.45 0.473
4 CLEAN + gradient 0.420 17.12 0.456
Spec Ranges (4 parameters)
Parameter Min Max Unit
baseline_error -0.0014 0.0046 m
phase_calibration -1.4 4.6 deg
amplitude_calibration 0.986 1.046 -
clock_offset -0.14 0.46 ns
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

F → S → D

F Fourier
S Sampling
D Detector

Mismatch Parameters

Symbol Parameter Description Nominal Perturbed
b_e baseline_error Baseline error (m) 0.0 0.002
p_c phase_calibration Phase calibration (deg) 0.0 2.0
a_c amplitude_calibration Amplitude calibration (-) 1.0 1.02
c_o clock_offset Clock offset (ns) 0.0 0.2

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.