Public

fMRI — 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
B0_inhomog -2.0 – 4.0 1.0 ppm
head_motion -1.0 – 2.0 0.5 mm
hemodynamic_delay 5.0 – 8.0 6.5 s
physiological_noise -0.02 – 0.04 0.01

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

17.61 dB

SSIM 0.3262

Scenario II (Mismatch)

13.04 dB

SSIM 0.0468

Scenario III (Oracle)

16.34 dB

SSIM 0.1224

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 16.58 0.3251 12.51 0.0447 15.93 0.1214
scene_01 18.88 0.3252 13.74 0.0480 17.02 0.1162
scene_02 18.30 0.3298 13.37 0.0496 16.55 0.1259
scene_03 16.68 0.3248 12.53 0.0450 15.85 0.1262

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 MR-IPT + gradient 0.892 40.84 0.99 0.94 ✓ Certified Sci. Reports 2025
2 SwinMR++ + gradient 0.889 41.48 0.991 0.89 ✓ Certified Huang et al., IEEE TMI 2025 — multi-scale axial attention + INR head + k-space DC per module + LPIPS+SSIM+k-space joint loss + dynamic feature fusion
3 MoDL-Net++ + gradient 0.888 40.79 0.99 0.92 ✓ Certified MoDL-Net++ IEEE TMI 2025 — multi-scale pyramid fusion + RDN/Swin deep prior + differentiable DC layers + LPIPS+SSIM+L1 joint loss + two-stage training strategy
4 ReconFormer++ + gradient 0.882 40.12 0.988 0.93 ✓ Certified Pan et al., IEEE TMI 2025
5 HUMUS-Net++ + gradient 0.880 40.63 0.989 0.89 ✓ Certified Fabian et al., dHUMUS-Net 2023 — k-space DC per module + dynamic multi-scale weighting + INR head + perceptual-structural loss + axial attention
6 PromptMR-SFM + gradient 0.880 39.73 0.987 0.94 ✓ Certified PWM 2026
7 PnP-DnCNN-Pro + gradient 0.878 39.54 0.987 0.94 ✓ Certified PnP-DnCNN-Pro IEEE TMI 2025 (DOI:10.1109/TMI.2025.3441240) — multi-scale DnCNN denoiser + adaptive mu/sigma schedule + SIREN INR output head + joint LPIPS+SSIM denoiser training + dynamic PnP regularization scheduling
8 BrainID-MRI + gradient 0.878 39.51 0.987 0.94 ✓ Certified Liu et al., CVPR 2025
9 HybridCascade++ + gradient 0.873 40.02 0.988 0.89 ✓ Certified HybridCascade++ MICCAI 2021 + IEEE TMI 2025 — multi-scale cascade DC + SIREN INR warm-start + SSIM structural anchor + DRUNet polish + freq-blend LF/HF fusion
10 MRI-FM + gradient 0.869 39.25 0.986 0.91 ✓ Certified Wang et al., Nature MI 2026
11 U-Net++ + gradient 0.864 39.51 0.987 0.87 ✓ Certified Chen & Boning, IEEE TMI 2024 (DOI: 10.1109/TMI.2024.3367890) — Residual U-Net + data consistency layers + plug-and-play prior + residual connections + dense skip paths
12 MMR-Mamba + gradient 0.858 38.83 0.985 0.88 ✓ Certified Zhao et al., Med. Image Anal. 2025
13 MRDynamo + gradient 0.852 38.23 0.983 0.88 ✓ Certified Chen et al., NeurIPS 2024
14 MRI-DiffusionNet + gradient 0.849 37.98 0.982 0.88 ✓ Certified Song et al., ICCV 2024
15 SwinMR + gradient 0.848 37.07 0.978 0.93 ✓ Certified Huang et al., MICCAI 2022
16 PromptMR + gradient 0.842 37.24 0.979 0.89 ✓ Certified Bai et al., ECCV 2024
17 E2E-VarNet + gradient 0.839 36.81 0.977 0.9 ✓ Certified Sriram et al., MICCAI 2020
18 HUMUS-Net + gradient 0.833 36.45 0.976 0.89 ✓ Certified Fabian et al., NeurIPS 2022
19 ReconFormer + gradient 0.833 36.15 0.974 0.91 ✓ Certified Guo et al., IEEE TMI 2024
20 HybridCascade + gradient 0.818 35.1 0.968 0.9 ✓ Certified Fastmri, arXiv 2020
21 U-Net + gradient 0.817 34.52 0.965 0.93 ✓ Certified Zbontar et al., arXiv 2018
22 Deep-ADMM-Net + gradient 0.808 33.69 0.958 0.94 ✓ Certified Yang et al., NeurIPS 2016
23 PnP-DnCNN + gradient 0.806 33.67 0.958 0.93 ✓ Certified Ahmad et al., IEEE TCI 2019 (DOI:10.1109/TCI.2019.2944521)
24 MoDL + gradient 0.803 33.9 0.96 0.9 ✓ Certified Aggarwal et al., IEEE TMI 2019
25 BM3D-MRI + gradient 0.794 32.66 0.949 0.94 ✓ Certified Eksioglu, Comput. Math. Meth. Med. 2016
26 DCCNN + gradient 0.788 32.76 0.95 0.9 ✓ Certified Schlemper et al., IEEE TMI 2018
27 ESPIRiT + gradient 0.782 31.8 0.94 0.94 ✓ Certified Uecker et al., MRM 2014
28 ALOHA + gradient 0.778 32.52 0.948 0.87 ✓ Certified Jin et al., IEEE TMI 2016
29 LORAKS + gradient 0.767 31.61 0.938 0.88 ✓ Certified Haldar, IEEE TMI 2014
30 L1-Wavelet + gradient 0.765 30.65 0.926 0.94 ✓ Certified Lustig et al., MRM 2007
31 k-t SPARSE-SENSE + gradient 0.743 29.78 0.913 0.9 ✓ Certified Lustig et al., MRM 2006
32 Score-MRI + gradient 0.728 29.12 0.902 0.88 ✓ Certified Chung & Ye, Med. Image Anal. 2022
33 SENSE + gradient 0.725 28.5 0.891 0.92 ✓ Certified Pruessmann et al., MRM 1999
34 GRAPPA + gradient 0.720 28.44 0.89 0.9 ✓ Certified Griswold et al., MRM 2002
35 Zero-Filled IFFT + gradient 0.624 24.2 0.776 0.86 ✓ Certified Pruessmann et al., MRM 1999

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 fMRI