Dev

Polarization — Dev Tier

(5 scenes)

Blind evaluation tier — no ground truth available.

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.

Parameter Specifications

🔒

True spec hidden — estimate parameters from spec ranges below.

Parameter Spec Range Unit
extinction_ratio -0.6 – 0.9 dB
retardance -2.4 – 3.6 nm
alignment -0.6 – 0.9 deg

Dev Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 Restormer+ + gradient 0.776 33.1 0.953 0.82 ✓ Certified Zamir et al., ICCV 2024
2 DeconvFormer + gradient 0.749 30.64 0.926 0.86 ✓ Certified Chen et al., CVPR 2024
3 ScoreMicro + gradient 0.746 31.35 0.935 0.79 ✓ Certified Wei et al., ECCV 2025
4 Restormer + gradient 0.729 29.99 0.917 0.81 ✓ Certified Zamir et al., CVPR 2022
5 ResUNet + gradient 0.718 29.36 0.907 0.81 ✓ Certified DeCelle et al., Nat. Methods 2021
6 DiffDeconv + gradient 0.697 27.51 0.87 0.87 ✓ Certified Huang et al., NeurIPS 2024
7 U-Net + gradient 0.670 25.88 0.829 0.9 ✓ Certified Ronneberger et al., MICCAI 2015
8 Wiener Filter + gradient 0.670 26.54 0.847 0.83 ✓ Certified Analytical baseline
9 CARE + gradient 0.665 25.81 0.827 0.88 ✓ Certified Weigert et al., Nat. Methods 2018
10 PnP-FISTA + gradient 0.659 25.71 0.824 0.86 ✓ Certified Bai et al., 2020
11 TV-Deconvolution + gradient 0.654 25.51 0.818 0.86 ✓ Certified Rudin et al., Phys. A 1992
12 PnP-DnCNN + gradient 0.652 25.85 0.828 0.81 ✓ Certified Zhang et al., IEEE TIP 2017
13 Richardson-Lucy + gradient 0.634 24.79 0.795 0.84 ✓ Certified Richardson, JOSA 1972 / Lucy, AJ 1974

Visible Data Fields

y H_ideal spec_ranges

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 Polarization