Public

OCT — Public Tier

(3 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
dispersion -200.0 – 400.0 100.0 fs²
reference_delay -5.0 – 10.0 2.5 μm
spectral_roll_off -1.0 – 2.0 0.5 dB/mm

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

24.44 dB

SSIM 0.8201

Scenario II (Mismatch)

19.91 dB

SSIM 0.6769

Scenario III (Oracle)

18.93 dB

SSIM 0.6208

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 25.17 0.8356 20.54 0.6765 19.60 0.6342
scene_01 24.16 0.8259 20.05 0.6783 18.86 0.6294
scene_02 22.70 0.7592 18.35 0.6361 17.48 0.5424
scene_03 25.71 0.8597 20.69 0.7166 19.79 0.6770

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 ScoreOCT + gradient 0.843 36.51 0.976 0.94 ✓ Certified Wei et al., ECCV 2025
2 DiffusionOCT + gradient 0.837 36.33 0.975 0.92 ✓ Certified Zhang et al., NeurIPS 2024
3 RetinalFormer + gradient 0.824 34.92 0.967 0.94 ✓ Certified Chen et al., ICCV 2024
4 OCT-ViT + gradient 0.819 34.39 0.964 0.95 ✓ Certified Tian et al., ICCV 2024
5 SpeckleFormer + gradient 0.806 34.31 0.963 0.89 ✓ Certified Devalla et al., ECCV 2024
6 U-Net-OCT + gradient 0.792 32.8 0.951 0.92 ✓ Certified Ronneberger et al., MICCAI 2015 (OCT variant)
7 OCTA-Net + gradient 0.777 32.09 0.944 0.89 ✓ Certified Hybrid U-Net+Transformer, 2023
8 Speckle-DenoiseNet + gradient 0.752 30.11 0.918 0.92 ✓ Certified Devalla et al., BOE 2019
9 NLM-OCT + gradient 0.730 28.44 0.89 0.95 ✓ Certified Buades et al., Multiscale Model. Simul. 2005
10 TV-Denoising + gradient 0.706 27.48 0.869 0.92 ✓ Certified Rudin et al., Phys. A 1992
11 BM4D + gradient 0.687 26.8 0.853 0.89 ✓ Certified Maggioni et al., IEEE TIP 2013
12 Speckle-Lee + gradient 0.655 25.19 0.808 0.9 ✓ Certified Lee, IEEE TGRS 1980
13 FFT-OCT + gradient 0.612 23.61 0.754 0.87 ✓ Certified Analytical baseline

Visible Data Fields

y H_ideal spec_ranges x_true true_spec

Dataset

Format: HDF5
Scenes: 3

Scoring Formula

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

PSNR: 40% SSIM: 40% Consistency: 20%
Back to OCT