Public

FPM — 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
led_position -0.1 – 0.2 0.05 mm
na_error 0.095 – 0.11 0.1025
defocus -2.0 – 4.0 1.0 μm

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

13.27 dB

SSIM 0.3991

Scenario II (Mismatch)

11.43 dB

SSIM 0.2034

Scenario III (Oracle)

18.77 dB

SSIM 0.3956

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 17.02 0.3838 14.01 0.2044 18.91 0.4324
scene_01 14.23 0.2969 12.52 0.1837 18.37 0.4876
scene_02 8.52 0.3871 7.90 0.1976 19.17 0.2761
scene_03 13.30 0.5284 11.31 0.2281 18.63 0.3860

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 PtychoDV + gradient 0.766 31.38 0.936 0.89 ✓ Certified Shamshad et al., IEEE TCI 2019
2 Fourier PtychoNet + gradient 0.741 29.65 0.911 0.9 ✓ Certified Jiang et al., BOE 2018
3 Gradient Descent FPM + gradient 0.666 25.51 0.818 0.92 ✓ Certified Tian & Waller, Optica 2015
4 Alternating Projections + gradient 0.620 23.29 0.742 0.95 ✓ Certified Zheng et al., Nat. Photonics 2013

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 FPM