Public

SPC-Block — Public Tier

(11 scenes)

Full-access development tier with all data visible.

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.

Parameter Specifications

True spec visible — use these exact values for Scenario III oracle reconstruction.

Parameter Spec Range True Value Unit
gain_decay_alpha 0.0005 – 0.0095 0.005 1/measurement
noise_sigma 0.01 – 0.05 0.03

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

25.18 dB

SSIM 0.7978

Scenario II (Mismatch)

18.31 dB

SSIM 0.7399

Scenario III (Oracle)

26.32 dB

SSIM 0.8133

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 25.39 0.8481 18.82 0.7840 26.51 0.8619
scene_01 27.01 0.8493 18.37 0.7963 26.52 0.8634
scene_02 21.78 0.6939 17.99 0.6610 24.86 0.7311
scene_03 26.51 0.8001 18.06 0.7181 27.40 0.7971

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 HATNet + gradient 0.710 27.91 0.778 0.88 ✓ Certified InverseNet baseline
2 ISTA-Net + gradient 0.694 26.61 0.742 0.92 ✓ Certified InverseNet baseline
3 FISTA-TV + gradient 0.680 25.29 0.738 0.91 ✓ Certified InverseNet baseline
4 PnP-DRUNet + gradient 0.627 22.99 0.643 0.93 ✓ Certified InverseNet baseline

Visible Data Fields

y H_ideal spec_ranges x_true true_spec

Dataset

Format: HDF5
Scenes: 11

Scoring Formula

0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)

PSNR: 40% SSIM: 40% Consistency: 20%
Back to SPC-Block