Dev

Matrix — 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
matrix_perturb -0.012 – 0.018
gain 0.964 – 1.054
sigma_y -0.024 – 0.036

Dev Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 ScoreSCI + gradient 0.751 30.77 0.928 0.86 ✓ Certified Chen et al., NeurIPS 2024
2 FlowHSI + gradient 0.745 30.29 0.921 0.87 ✓ Certified Huang et al., arXiv 2025
3 CSTrans + gradient 0.744 30.98 0.931 0.81 ✓ Certified Liu et al., CVPR 2024
4 CST + gradient 0.742 29.82 0.914 0.89 ✓ Certified Liu et al., ICCV 2023
5 Restormer + gradient 0.739 30.37 0.922 0.83 ✓ Certified Zamir et al., CVPR 2022
6 MST-L + gradient 0.736 30.72 0.927 0.79 ✓ Certified Cai et al., CVPR 2022
7 PromptSCI + gradient 0.731 30.42 0.923 0.79 ✓ Certified Bai et al., ICCV 2024
8 HiSViT+ + gradient 0.728 29.12 0.902 0.88 ✓ Certified Tao et al., ECCV 2024
9 DiffusionHSI + gradient 0.703 28.57 0.892 0.8 ✓ Certified Zhang et al., ICCV 2024
10 EfficientSCI + gradient 0.677 27.21 0.863 0.8 ✓ Certified Wang et al., IEEE TIP 2023
11 TVAL3 + gradient 0.656 25.41 0.815 0.88 ✓ Certified Li et al., SIAM J. Sci. Comput. 2009
12 FISTA-TV + gradient 0.642 24.67 0.792 0.89 ✓ Certified Beck & Teboulle, SIAM J. Imaging Sci. 2009
13 GAP-TV + gradient 0.627 24.47 0.785 0.84 ✓ Certified Yuan et al., IEEE TIP 2016
14 PnP-FFDNet + gradient 0.583 22.58 0.714 0.86 ✓ Certified Zhang et al., IEEE TPAMI 2020

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 Matrix