Public

CBCT — Public Tier

(10 scenes)

Full-access tier: 10 real CBCT/CT volumes from AAPM, LIDC-IDRI, CBCTLiTS, MMDental, CTooth+, 2DeteCT, HTC, Walnut CT, CQ500, DM4CT.

What you get

Cone-beam projections (y), ideal geometry (H), spec ranges, ground truth volume (x_true), and true mismatch spec.

How to use

Load cbct_challenge_public.h5 → reconstruct 256³ volume from projections → compare with x_true → iterate on mismatch correction.

What to submit

Reconstructed volumes (x_hat) and corrected mismatch spec as HDF5.

Parameter Specifications

✓

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

Parameter Spec Range True Value Unit
source_offset_x -1.2 – 2.8 0.8 mm
source_offset_z -1.0 – 2.0 0.5 mm
detector_tilt -0.35 – 0.65 0.15 deg
detector_shift_u -1.8 – 4.2 1.2 px
beam_hardening -0.015 – 0.135 0.06
scatter_fraction -0.01 – 0.09 0.04

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 DiffusionCBCT + gradient 0.848 37.92 0.982 0.88 ✓ Certified Gao et al., Med. Phys. 2024
2 CTFormer + gradient 0.835 36.97 0.978 0.87 ✓ Certified Wang et al., MICCAI 2023
3 DuDoTrans + gradient 0.823 35.31 0.97 0.91 ✓ Certified Wang et al., IEEE TMI 2022
4 Metal-AR-Net + gradient 0.818 34.72 0.966 0.92 ✓ Certified Zhang & Yu, IEEE TMI 2018
5 DuDoNet + gradient 0.811 34.53 0.965 0.9 ✓ Certified Lin et al., CVPR 2019
6 Learned Primal-Dual + gradient 0.804 34.48 0.964 0.87 ✓ Certified Adler & Oktem, IEEE TMI 2018
7 FBPConvNet + gradient 0.774 31.78 0.94 0.9 ✓ Certified Jin et al., IEEE TIP 2017
8 TV-ADMM + gradient 0.749 29.63 0.911 0.94 ✓ Certified Boyd et al., Found. Trends 2011
9 FDK + gradient 0.685 26.2 0.838 0.94 ✓ Certified Feldkamp et al., J. Opt. Soc. Am. A 1984

Visible Data Fields

y H_ideal spec_ranges x_true true_spec

Dataset

Format: HDF5
Scenes: 10

Scoring Formula

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

PSNR: 40% SSIM: 40% Consistency: 20%
Back to CBCT