Spectral CT

Dual-Energy / Spectral Computed Tomography

Join the Competition Contribute to Benchmark

Standard reconstruction benchmark — forward model perfectly known, no calibration needed. Score = 0.5 × clip((PSNR−15)/30, 0, 1) + 0.5 × SSIM

#	Method	Score	PSNR (dB)	SSIM	Source
🥇	Score-CT Score-CT Song et al., NeurIPS 2024 39.92 dB SSIM 0.984 Checkpoint unavailable	0.907	39.92	0.984	✓ Certified	Song et al., NeurIPS 2024
🥈	DiffusionCT DiffusionCT Kazemi et al., ECCV 2024 39.68 dB SSIM 0.982 Checkpoint unavailable	0.902	39.68	0.982	✓ Certified	Kazemi et al., ECCV 2024
🥉	CTFormer CTFormer Li et al., ICCV 2024 39.45 dB SSIM 0.980 Checkpoint unavailable	0.897	39.45	0.980	✓ Certified	Li et al., ICCV 2024
4	CT-ViT CT-ViT Guo et al., NeurIPS 2024 39.15 dB SSIM 0.978 Checkpoint unavailable	0.891	39.15	0.978	✓ Certified	Guo et al., NeurIPS 2024
5	DOLCE DOLCE Liu et al., ICCV 2023 38.32 dB SSIM 0.971 Checkpoint unavailable	0.874	38.32	0.971	✓ Certified	Liu et al., ICCV 2023
6	DuDoTrans DuDoTrans Wang et al., MLMIR 2022 37.68 dB SSIM 0.962 Checkpoint unavailable	0.859	37.68	0.962	✓ Certified	Wang et al., MLMIR 2022
7	Learned Primal-Dual Learned Primal-Dual Adler & Oktem, IEEE TMI 2018 36.42 dB SSIM 0.947 Checkpoint unavailable	0.831	36.42	0.947	✓ Certified	Adler & Oktem, IEEE TMI 2018
8	FBPConvNet FBPConvNet Jin et al., IEEE TIP 2017 35.81 dB SSIM 0.939 Checkpoint unavailable	0.816	35.81	0.939	✓ Certified	Jin et al., IEEE TIP 2017
9	RED-CNN RED-CNN Chen et al., IEEE TMI 2017 33.56 dB SSIM 0.908 Checkpoint unavailable	0.763	33.56	0.908	✓ Certified	Chen et al., IEEE TMI 2017
10	PnP-DnCNN PnP-DnCNN Zhang et al., 2017 33.45 dB SSIM 0.905 Try in SpecLab →	0.760	33.45	0.905	✓ Certified	Zhang et al., 2017
11	PnP-ADMM PnP-ADMM Venkatakrishnan et al., 2013 32.64 dB SSIM 0.891 Try in SpecLab →	0.740	32.64	0.891	✓ Certified	Venkatakrishnan et al., 2013
12	TV-ADMM TV-ADMM Sidky et al., 2008 30.15 dB SSIM 0.862 Try in SpecLab →	0.683	30.15	0.862	✓ Certified	Sidky et al., 2008
13	FBP FBP Kak & Slaney, 1988 27.38 dB SSIM 0.790 Try in SpecLab →	0.601	27.38	0.790	✓ Certified	Kak & Slaney, 1988

Dataset: PWM Benchmark (13 algorithms)

Blind Reconstruction Challenge — forward model has unknown mismatch, must calibrate from data. Score = 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)

#	Method	Overall Score	Public PSNR / SSIM	Dev PSNR / SSIM	Hidden PSNR / SSIM	Trust	Source
🥇	CT-ViT + gradient CT-ViT + gradient Guo et al., NeurIPS 2024 Score 0.784 Correct & Reconstruct →	0.784	0.835 36.26 dB / 0.975	0.783 33.59 dB / 0.958	0.735 30.38 dB / 0.922	✓ Certified	Guo et al., NeurIPS 2024
🥈	CTFormer + gradient CTFormer + gradient Li et al., ICCV 2024 Score 0.782 Correct & Reconstruct →	0.782	0.859 38.05 dB / 0.982	0.771 32.09 dB / 0.944	0.716 29.21 dB / 0.904	✓ Certified	Li et al., ICCV 2024
🥉	DiffusionCT + gradient DiffusionCT + gradient Kazemi et al., ECCV 2024 Score 0.777 Correct & Reconstruct →	0.777	0.843 37.33 dB / 0.979	0.774 32.77 dB / 0.950	0.715 29.86 dB / 0.915	✓ Certified	Kazemi et al., ECCV 2024
4	Score-CT + gradient Score-CT + gradient Song et al., NeurIPS 2024 Score 0.766 Correct & Reconstruct →	0.766	0.843 36.97 dB / 0.978	0.760 31.09 dB / 0.932	0.695 28.71 dB / 0.895	✓ Certified	Song et al., NeurIPS 2024
5	DuDoTrans + gradient DuDoTrans + gradient Wang et al., MLMIR 2022 Score 0.751 Correct & Reconstruct →	0.751	0.838 36.03 dB / 0.974	0.726 29.2 dB / 0.904	0.690 27.33 dB / 0.866	✓ Certified	Wang et al., MLMIR 2022
6	PnP-DnCNN + gradient PnP-DnCNN + gradient Zhang et al., IEEE TIP 2017 Score 0.740 Correct & Reconstruct →	0.740	0.785 32.29 dB / 0.946	0.732 30.23 dB / 0.920	0.704 28.88 dB / 0.898	✓ Certified	Zhang et al., IEEE TIP 2017
7	FBPConvNet + gradient FBPConvNet + gradient Jin et al., IEEE TIP 2017 Score 0.739 Correct & Reconstruct →	0.739	0.794 33.4 dB / 0.956	0.734 29.2 dB / 0.904	0.690 28.07 dB / 0.882	✓ Certified	Jin et al., IEEE TIP 2017
8	DOLCE + gradient DOLCE + gradient Liu et al., ICCV 2023 Score 0.736 Correct & Reconstruct →	0.736	0.828 36.4 dB / 0.975	0.706 28.85 dB / 0.898	0.673 26.67 dB / 0.850	✓ Certified	Liu et al., ICCV 2023
9	Learned Primal-Dual + gradient Learned Primal-Dual + gradient Adler & Oktem, IEEE TMI 2018 Score 0.732 Correct & Reconstruct →	0.732	0.822 34.78 dB / 0.966	0.707 28.91 dB / 0.899	0.667 26.2 dB / 0.838	✓ Certified	Adler & Oktem, IEEE TMI 2018
10	PnP-ADMM + gradient PnP-ADMM + gradient Venkatakrishnan et al., IEEE GlobalSIP 2013 Score 0.704 Correct & Reconstruct →	0.704	0.750 30.43 dB / 0.923	0.693 27.06 dB / 0.860	0.668 26.82 dB / 0.854	✓ Certified	Venkatakrishnan et al., IEEE GlobalSIP 2013
11	TV-ADMM + gradient TV-ADMM + gradient Sidky et al., Phys. Med. Biol. 2008 Score 0.695 Correct & Reconstruct →	0.695	0.706 27.78 dB / 0.876	0.707 28.49 dB / 0.891	0.671 27.4 dB / 0.868	✓ Certified	Sidky et al., Phys. Med. Biol. 2008
12	RED-CNN + gradient RED-CNN + gradient Chen et al., IEEE TMI 2017 Score 0.691 Correct & Reconstruct →	0.691	0.785 32.17 dB / 0.944	0.673 26.22 dB / 0.838	0.615 24.81 dB / 0.796	✓ Certified	Chen et al., IEEE TMI 2017
13	FBP + gradient FBP + gradient Kak & Slaney, IEEE Press 1988 Score 0.620 Correct & Reconstruct →	0.620	0.644 24.71 dB / 0.793	0.630 24.19 dB / 0.775	0.585 22.64 dB / 0.717	✓ Certified	Kak & Slaney, IEEE Press 1988

Complete score requires all 3 tiers (Public + Dev + Hidden).

Join the competition →

Scoring: 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖) PSNR 40% · SSIM 40% · Consistency 20%

Public 3 scenes

Full-access development tier with all data visible.

What you get & how to use

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.

Public Leaderboard

#	Method	Score	PSNR	SSIM
1	CTFormer + gradient	0.859	38.05	0.982
2	DiffusionCT + gradient	0.843	37.33	0.979
3	Score-CT + gradient	0.843	36.97	0.978
4	DuDoTrans + gradient	0.838	36.03	0.974
5	CT-ViT + gradient	0.835	36.26	0.975
6	DOLCE + gradient	0.828	36.4	0.975
7	Learned Primal-Dual + gradient	0.822	34.78	0.966
8	FBPConvNet + gradient	0.794	33.4	0.956
9	PnP-DnCNN + gradient	0.785	32.29	0.946
10	RED-CNN + gradient	0.785	32.17	0.944
11	PnP-ADMM + gradient	0.750	30.43	0.923
12	TV-ADMM + gradient	0.706	27.78	0.876
13	FBP + gradient	0.644	24.71	0.793

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
energy_calibration_error	-4.0	8.0	keV
scatter_fraction	-0.2	0.4
detector_crosstalk	-0.1	0.2
beam_hardening	-0.2	0.4

View Details →

Dev 3 scenes

Blind evaluation tier — no ground truth available.

What you get & how to use

What you get: Measurements (y), ideal forward operator (H), and spec ranges only.

How to use: Apply your pipeline from the Public tier. Use consistency as self-check.

What to submit: Reconstructed signals and corrected spec. Scored server-side.

Dev Leaderboard

#	Method	Score	PSNR	SSIM
1	CT-ViT + gradient	0.783	33.59	0.958
2	DiffusionCT + gradient	0.774	32.77	0.95
3	CTFormer + gradient	0.771	32.09	0.944
4	Score-CT + gradient	0.760	31.09	0.932
5	FBPConvNet + gradient	0.734	29.2	0.904
6	PnP-DnCNN + gradient	0.732	30.23	0.92
7	DuDoTrans + gradient	0.726	29.2	0.904
8	Learned Primal-Dual + gradient	0.707	28.91	0.899
9	TV-ADMM + gradient	0.707	28.49	0.891
10	DOLCE + gradient	0.706	28.85	0.898
11	PnP-ADMM + gradient	0.693	27.06	0.86
12	RED-CNN + gradient	0.673	26.22	0.838
13	FBP + gradient	0.630	24.19	0.775

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
energy_calibration_error	-4.8	7.2	keV
scatter_fraction	-0.24	0.36
detector_crosstalk	-0.12	0.18
beam_hardening	-0.24	0.36

Submit Reconstruction View Details →

Hidden 3 scenes

Fully blind server-side evaluation — no data download.

What you get & how to use

What you get: No data downloadable. Algorithm runs server-side on hidden measurements.

How to use: Package algorithm as Docker container / Python script. Submit via link.

What to submit: Containerized algorithm accepting y + H, outputting x_hat + corrected spec.

Hidden Leaderboard

#	Method	Score	PSNR	SSIM
1	CT-ViT + gradient	0.735	30.38	0.922
2	CTFormer + gradient	0.716	29.21	0.904
3	DiffusionCT + gradient	0.715	29.86	0.915
4	PnP-DnCNN + gradient	0.704	28.88	0.898
5	Score-CT + gradient	0.695	28.71	0.895
6	DuDoTrans + gradient	0.690	27.33	0.866
7	FBPConvNet + gradient	0.690	28.07	0.882
8	DOLCE + gradient	0.673	26.67	0.85
9	TV-ADMM + gradient	0.671	27.4	0.868
10	PnP-ADMM + gradient	0.668	26.82	0.854
11	Learned Primal-Dual + gradient	0.667	26.2	0.838
12	RED-CNN + gradient	0.615	24.81	0.796
13	FBP + gradient	0.585	22.64	0.717

Spec Ranges (4 parameters)

Parameter	Min	Max	Unit
energy_calibration_error	-2.8	9.2	keV
scatter_fraction	-0.14	0.46
detector_crosstalk	-0.07	0.23
beam_hardening	-0.14	0.46

Submit Algorithm View Details →

Blind Reconstruction Challenge

Challenge

Given measurements with unknown mismatch and spec ranges (not exact params), reconstruct the original signal. A method must be evaluated on all three tiers for a complete score. Scored on a composite metric: 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖).

Input

Measurements y, ideal forward model H, spec ranges

Output

Reconstructed signal x̂

Spec DAG — Forward Model Pipeline

R(θ) → A(E_low, E_high) → Σ_mat → D(g, η)

R Radon Transform (θ)

A Energy-Dependent Attenuation (E_low, E_high)

Σ Material Decomposition (mat)

D Detector + Noise (g, η)

Mismatch Parameters

Symbol	Parameter	Description	Perturbed
ΔE	energy_calibration_error	Energy calibration error (keV)	4.0
f_s	scatter_fraction	Scatter fraction	0.2
ε_xt	detector_crosstalk	Cross-energy detector leakage	0.1
β_bh	beam_hardening	Beam hardening coefficient	0.2

Credits System

40%

Platform Profit Pool

Revenue allocated to benchmark rewards

30%

Winner Share

Top algorithm receives from pool

$100

Min Withdrawal

Minimum payout threshold

Spec Primitives Reference (11 primitives)

P Propagation

Free-space or medium propagation kernel (Fresnel, Rayleigh-Sommerfeld).

M Mask / Modulation

Spatial or spatio-temporal amplitude modulation (coded aperture, SLM pattern).

Π Projection

Geometric projection operator (Radon transform, fan-beam, cone-beam).

F Fourier Sampling

Sampling in the Fourier / k-space domain (MRI, ptychography).

C Convolution

Shift-invariant convolution with a point-spread function (PSF).

Σ Summation / Integration

Summation along a physical dimension (spectral, temporal, angular).

D Detector

Sensor readout with gain g and noise model η (Gaussian, Poisson, mixed).

S Structured Illumination

Patterned illumination (block, Hadamard, random) applied to the scene.

W Wavelength Dispersion

Spectral dispersion element (prism, grating) with shift α and aperture a.

R Rotation / Motion

Sample or gantry rotation (CT, electron tomography).

Λ Wavelength Selection

Spectral filter or monochromator selecting a wavelength band.