Public

Light-Sheet — Public Tier

(5 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
sheet_thickness -1.0 – 2.0 0.5 μm
sheet_tilt -0.5 – 1.0 0.25 deg
stripe_artifact -0.1 – 0.2 0.05

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

21.71 dB

SSIM 0.6116

Scenario II (Mismatch)

34.59 dB

SSIM 0.8586

Scenario III (Oracle)

7.20 dB

SSIM 0.0058

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 20.03 0.6745 41.11 0.9298 8.59 0.0045
scene_01 27.62 0.8195 32.02 0.9127 8.64 0.0014
scene_02 12.96 0.2219 31.09 0.6569 5.87 0.0102
scene_03 26.20 0.7305 34.13 0.9348 5.68 0.0069

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 ScoreMicro + gradient 0.848 36.73 0.977 0.95 ✓ Certified Wei et al., ECCV 2025
2 Restormer+ + gradient 0.838 36.04 0.974 0.94 ✓ Certified Zamir et al., ICCV 2024
3 DeconvFormer + gradient 0.834 36.06 0.974 0.92 ✓ Certified Chen et al., CVPR 2024
4 DiffDeconv + gradient 0.822 35.58 0.971 0.89 ✓ Certified Huang et al., NeurIPS 2024
5 Restormer + gradient 0.818 34.77 0.966 0.92 ✓ Certified Zamir et al., CVPR 2022
6 U-Net + gradient 0.806 33.52 0.957 0.94 ✓ Certified Ronneberger et al., MICCAI 2015
7 ResUNet + gradient 0.794 33.36 0.956 0.89 ✓ Certified DeCelle et al., Nat. Methods 2021
8 CARE + gradient 0.778 32.35 0.946 0.88 ✓ Certified Weigert et al., Nat. Methods 2018
9 PnP-DnCNN + gradient 0.748 29.56 0.91 0.94 ✓ Certified Zhang et al., IEEE TIP 2017
10 PnP-FISTA + gradient 0.736 28.75 0.896 0.95 ✓ Certified Bai et al., 2020
11 TV-Deconvolution + gradient 0.719 27.81 0.877 0.95 ✓ Certified Rudin et al., Phys. A 1992
12 Wiener Filter + gradient 0.662 25.43 0.815 0.91 ✓ Certified Analytical baseline
13 Richardson-Lucy + gradient 0.650 25.3 0.812 0.86 ✓ Certified Richardson, JOSA 1972 / Lucy, AJ 1974

Visible Data Fields

y H_ideal spec_ranges x_true true_spec

Dataset

Format: HDF5
Scenes: 5

Scoring Formula

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

PSNR: 40% SSIM: 40% Consistency: 20%
Back to Light-Sheet