Public

Magnetic Particle Imaging (MPI) — Public Tier

(5 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
drive_field_amplitude	24.4 – 26.2	25.3	mT
selection_field_gradient	2.4 – 2.7	2.55	T/m
particle_relaxation_time	1.8 – 2.4	2.1	us
receive_coil_sensitivity	0.97 – 1.06	1.015	-

Method: CPU_baseline — Mismatch parameter: nominal

7.77 dB

SSIM 0.3772

7.61 dB

SSIM 0.1940

15.93 dB

SSIM 0.4222

Per-scene breakdown (4 scenes)

Scene	PSNR I	SSIM I	PSNR II	SSIM II	PSNR III	SSIM III
scene_00	7.57	0.3615	7.48	0.1918	15.88	0.4337
scene_01	7.70	0.3833	7.85	0.1983	15.89	0.4210
scene_02	7.97	0.3846	7.45	0.1929	16.02	0.4175
scene_03	7.84	0.3794	7.67	0.1931	15.94	0.4166

#	Method	Score	PSNR	SSIM	Consistency	Trust	Source
1	SwinIR + gradient	0.793	32.39	0.947	0.95	✓ Certified	Liang et al., ICCVW 2021
2	ResUNet + gradient	0.771	30.94	0.93	0.95	✓ Certified	Residual U-Net baseline
3	ExpFormer + gradient	0.766	30.55	0.925	0.95	✓ Certified	Experimental science transformer, 2024
4	ScoreExperimental + gradient	0.765	31.57	0.938	0.87	✓ Certified	Wei et al., 2025
5	Domain-Adapted-CNN + gradient	0.737	29.25	0.905	0.91	✓ Certified	Domain adaptation CNN
6	DiffusionExperimental + gradient	0.715	28.38	0.889	0.88	✓ Certified	Zhang et al., 2024
7	PnP-ADMM + gradient	0.694	27.02	0.859	0.9	✓ Certified	ADMM + denoiser prior
8	PnP-RED + gradient	0.686	27.04	0.859	0.86	✓ Certified	Romano et al., IEEE TIP 2017
9	Wiener Filter + gradient	0.675	25.64	0.822	0.95	✓ Certified	Wiener filtering baseline
10	Matched Filter + gradient	0.620	23.52	0.751	0.92	✓ Certified	Optimal linear filter
11	Tikhonov + gradient	0.602	23.14	0.737	0.88	✓ Certified	Tikhonov, Doklady 1963

y H_ideal spec_ranges x_true true_spec

Format: HDF5

Scenes: 5

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

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