Public

Portal Imaging (EPID) — 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
isocenter_shift -0.4 – 0.8 0.2 mm
beam_energy_variation 5.96 – 6.08 6.02 MV
detector_sag -0.2 – 0.4 0.1 mm
scatter_kernel_width 4.6 – 5.8 5.2 mm

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

14.88 dB

SSIM 0.2885

Scenario II (Mismatch)

11.40 dB

SSIM 0.0410

Scenario III (Oracle)

14.64 dB

SSIM 0.1177

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 14.80 0.2898 11.38 0.0400 14.58 0.1169
scene_01 15.05 0.2857 11.41 0.0409 14.66 0.1169
scene_02 14.99 0.2846 11.40 0.0427 14.66 0.1192
scene_03 14.69 0.2938 11.40 0.0405 14.65 0.1180

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 Score-CT + gradient 0.865 38.22 0.983 0.95 ✓ Certified Song et al., NeurIPS 2024
2 CT-ViT + gradient 0.857 37.86 0.982 0.93 ✓ Certified Guo et al., NeurIPS 2024
3 DiffusionCT + gradient 0.844 37.8 0.981 0.87 ✓ Certified Kazemi et al., ECCV 2024
4 CTFormer + gradient 0.839 36.63 0.976 0.91 ✓ Certified Li et al., ICCV 2024
5 DOLCE + gradient 0.828 36.42 0.975 0.87 ✓ Certified Liu et al., ICCV 2023
6 Learned Primal-Dual + gradient 0.824 34.93 0.967 0.94 ✓ Certified Adler & Oktem, IEEE TMI 2018
7 DuDoTrans + gradient 0.820 35.73 0.972 0.87 ✓ Certified Wang et al., MLMIR 2022
8 FBPConvNet + gradient 0.795 33.43 0.956 0.89 ✓ Certified Jin et al., IEEE TIP 2017
9 PnP-DnCNN + gradient 0.787 32.4 0.947 0.92 ✓ Certified Zhang et al., IEEE TIP 2017
10 PnP-ADMM + gradient 0.770 30.89 0.929 0.95 ✓ Certified Venkatakrishnan et al., IEEE GlobalSIP 2013
11 RED-CNN + gradient 0.763 31.34 0.935 0.88 ✓ Certified Chen et al., IEEE TMI 2017
12 TV-ADMM + gradient 0.730 28.53 0.891 0.94 ✓ Certified Sidky et al., Phys. Med. Biol. 2008
13 FBP + gradient 0.644 24.7 0.792 0.9 ✓ Certified Kak & Slaney, IEEE Press 1988

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 Portal Imaging (EPID)