Passive Microwave Radiometry

Passive Microwave Radiometry

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
🥇 RadioNet 0.751 31.81 0.941 ✓ Certified Passive microwave CNN, 2022
🥈 MWR-Former 0.727 30.78 0.928 ✓ Certified Microwave radiometry transformer, 2024
🥉 Backus-Gilbert 0.520 23.6 0.754 ✓ Certified Backus & Gilbert, Geophys. J. 1968
4 Tikhonov-SMOS 0.472 22.27 0.701 ✓ Certified Anterrieu, IEEE TGRS 2004

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
🥇 MWR-Former + gradient 0.656
0.717
28.38 dB / 0.889
0.673
26.61 dB / 0.848
0.578
23.06 dB / 0.733
✓ Certified Microwave radiometry transformer, 2024
🥈 RadioNet + gradient 0.623
0.735
29.53 dB / 0.909
0.591
22.79 dB / 0.723
0.542
21.11 dB / 0.651
✓ Certified Passive microwave CNN, 2022
🥉 Tikhonov-SMOS + gradient 0.510
0.514
19.87 dB / 0.592
0.509
19.7 dB / 0.584
0.506
20.22 dB / 0.609
✓ Certified Anterrieu, IEEE TGRS 2004
4 Backus-Gilbert + gradient 0.509
0.541
20.6 dB / 0.627
0.495
19.42 dB / 0.571
0.490
19.62 dB / 0.580
✓ Certified Backus & Gilbert, Geophys. J. 1968

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 RadioNet + gradient 0.735 29.53 0.909
2 MWR-Former + gradient 0.717 28.38 0.889
3 Backus-Gilbert + gradient 0.541 20.6 0.627
4 Tikhonov-SMOS + gradient 0.514 19.87 0.592
Spec Ranges (4 parameters)
Parameter Min Max Unit
antenna_beam_width_error -0.1 0.2 deg
receiver_gain_drift 0.99 1.02 -
brightness_temperature_offset -0.4 0.8 K
cross_polarization_leakage -0.004 0.008 -
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 MWR-Former + gradient 0.673 26.61 0.848
2 RadioNet + gradient 0.591 22.79 0.723
3 Tikhonov-SMOS + gradient 0.509 19.7 0.584
4 Backus-Gilbert + gradient 0.495 19.42 0.571
Spec Ranges (4 parameters)
Parameter Min Max Unit
antenna_beam_width_error -0.12 0.18 deg
receiver_gain_drift 0.988 1.018 -
brightness_temperature_offset -0.48 0.72 K
cross_polarization_leakage -0.0048 0.0072 -
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 MWR-Former + gradient 0.578 23.06 0.733
2 RadioNet + gradient 0.542 21.11 0.651
3 Tikhonov-SMOS + gradient 0.506 20.22 0.609
4 Backus-Gilbert + gradient 0.490 19.62 0.58
Spec Ranges (4 parameters)
Parameter Min Max Unit
antenna_beam_width_error -0.07 0.23 deg
receiver_gain_drift 0.993 1.023 -
brightness_temperature_offset -0.28 0.92 K
cross_polarization_leakage -0.0028 0.0092 -
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

Σ → D

Σ Summation
D Detector

Mismatch Parameters

Symbol Parameter Description Nominal Perturbed
a_b antenna_beam_width_error Antenna beam width error (deg) 0.0 0.1
r_g receiver_gain_drift Receiver gain drift (-) 1.0 1.01
b_t brightness_temperature_offset Brightness temperature offset (K) 0.0 0.4
c_l cross_polarization_leakage Cross-polarization leakage (-) 0.0 0.004

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.