Public

PET/MR — Public Tier

(3 scenes)

Full-access development tier with all data visible.

Measurements (y), ideal forward operator (H), spec ranges, ground truth (x_true), and true mismatch spec.

Load HDF5 → compare reconstruction vs x_true → check consistency → iterate.

Reconstructed signals (x_hat) and corrected spec as HDF5.

✓

True spec visible — use these exact values for Scenario III oracle reconstruction.

Parameter	Spec Range	True Value	Unit
mr_attenuation_error	-25.0 – 50.0	12.5	%
motion_shift	-6.0 – 12.0	3.0	pixels
b0_inhomogeneity	-40.0 – 80.0	20.0	Hz
pet_scatter_fraction	-0.25 – 0.5	0.125

Method: CPU_baseline — Mismatch parameter: nominal

7.72 dB

SSIM 0.3738

7.66 dB

SSIM 0.2292

16.33 dB

SSIM 0.4587

Per-scene breakdown (4 scenes)

Scene	PSNR I	SSIM I	PSNR II	SSIM II	PSNR III	SSIM III
scene_00	7.57	0.3615	7.52	0.2249	16.28	0.4660
scene_01	7.70	0.3833	7.81	0.2328	16.28	0.4522
scene_02	7.97	0.3846	7.55	0.2285	16.43	0.4500
scene_03	7.63	0.3655	7.75	0.2308	16.33	0.4667

#	Method	Score	PSNR	SSIM	Consistency	Trust	Source
1	PET-ViT + gradient	0.848	37.28	0.979	0.92	✓ Certified	Smith et al., ICCV 2024
2	PETFormer + gradient	0.780	32.61	0.949	0.87	✓ Certified	Li et al., ECCV 2024
3	U-Net-PET + gradient	0.779	32.38	0.947	0.88	✓ Certified	Ronneberger et al. variant, MICCAI 2020
4	TransEM + gradient	0.768	31.76	0.94	0.87	✓ Certified	Xie et al., 2023
5	DeepPET + gradient	0.741	29.4	0.907	0.92	✓ Certified	Haggstrom et al., MIA 2019
6	ML-EM + gradient	0.715	27.73	0.875	0.94	✓ Certified	Shepp & Vardi, IEEE TPAMI 1982
7	MAPEM-RDP + gradient	0.705	27.41	0.868	0.92	✓ Certified	Nuyts et al., IEEE TMI 2002
8	FBP-PET + gradient	0.703	27.6	0.872	0.89	✓ Certified	Analytical baseline
9	OS-EM + gradient	0.659	25.35	0.813	0.9	✓ Certified	Hudson & Larkin, IEEE TMI 1994
10	OSEM + gradient	0.622	23.63	0.755	0.92	✓ Certified	Hudson & Larkin, IEEE TMI 1994

y H_ideal spec_ranges x_true true_spec

Format: HDF5

Scenes: 3

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

PSNR: 40% SSIM: 40% Consistency: 20%