X-ray Angiography

X-ray Angiography

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
🥇 DiffusionAngio 0.847 36.8 0.967 ✓ Certified Shen et al., Med. Image Anal. 2024
🥈 AngioFormer 0.833 36.2 0.960 ✓ Certified Geometry-aware transformer 3DRA, 2024
🥉 NeRF-Angio 0.824 35.8 0.955 ✓ Certified Wang et al., IEEE TMI 43:1401, 2024
4 VesselNet 0.811 35.2 0.948 ✓ Certified Zhang et al., Radiology AI 2024
5 Learned Primal-Dual 0.792 34.5 0.935 ✓ Certified Adler & Oktem, IEEE TMI 2018
6 FBPConvNet 0.768 33.5 0.920 ✓ Certified Jin et al., IEEE TIP 2017
7 PnP-ADMM 0.730 32.0 0.893 ✓ Certified Venkatakrishnan et al., 2013
8 TV-CS 0.688 30.5 0.860 ✓ Certified Sidky et al., Phys. Med. Biol. 2008
9 FDK 0.590 27.0 0.780 ✓ Certified Feldkamp et al., JOSA A 1984

Dataset: PWM Benchmark (9 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
🥇 AngioFormer + gradient 0.741
0.799
33.43 dB / 0.956
0.746
30.2 dB / 0.920
0.677
27.62 dB / 0.873
✓ Certified Geometry-aware transformer for few-view 3DRA, 2024
🥈 NeRF-Angio + gradient 0.729
0.818
34.78 dB / 0.966
0.704
28.4 dB / 0.889
0.665
26.4 dB / 0.843
✓ Certified Wang et al., IEEE Trans. Med. Imaging 43:1401, 2024
🥉 DiffusionAngio + gradient 0.700
0.830
35.7 dB / 0.972
0.667
25.83 dB / 0.827
0.603
23.98 dB / 0.768
✓ Certified Shen et al., Med. Image Anal. 94:103102, 2024
4 TV-CS + gradient 0.697
0.716
28.65 dB / 0.894
0.698
27.99 dB / 0.881
0.678
26.94 dB / 0.857
✓ Certified Rudin et al., Physica D 60:259, 1992; Sidky et al., PMB 2008
5 VesselNet + gradient 0.691
0.809
33.92 dB / 0.960
0.669
26.61 dB / 0.848
0.594
23.62 dB / 0.755
✓ Certified Zhang et al., Radiology AI 6:e230298, 2024
6 Learned Primal-Dual + gradient 0.686
0.778
32.48 dB / 0.948
0.660
25.6 dB / 0.821
0.621
24.65 dB / 0.791
✓ Certified Adler & Oktem, IEEE TMI 37:1322, 2018
7 PnP-ADMM + gradient 0.682
0.742
30.07 dB / 0.918
0.666
26.04 dB / 0.833
0.639
24.85 dB / 0.797
✓ Certified Venkatakrishnan et al., IEEE GlobalSIP 2013
8 FBPConvNet + gradient 0.672
0.787
32.27 dB / 0.946
0.653
25.56 dB / 0.819
0.577
22.95 dB / 0.729
✓ Certified Jin et al., IEEE TIP 26:4509, 2017
9 FDK + gradient 0.605
0.646
25.05 dB / 0.804
0.604
24.0 dB / 0.769
0.565
22.11 dB / 0.695
✓ Certified Feldkamp et al., JOSA A 1(6):612, 1984

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 DiffusionAngio + gradient 0.830 35.7 0.972
2 NeRF-Angio + gradient 0.818 34.78 0.966
3 VesselNet + gradient 0.809 33.92 0.96
4 AngioFormer + gradient 0.799 33.43 0.956
5 FBPConvNet + gradient 0.787 32.27 0.946
6 Learned Primal-Dual + gradient 0.778 32.48 0.948
7 PnP-ADMM + gradient 0.742 30.07 0.918
8 TV-CS + gradient 0.716 28.65 0.894
9 FDK + gradient 0.646 25.05 0.804
Spec Ranges (3 parameters)
Parameter Min Max Unit
contrast_timing -0.5 1.0 s
motion -2.0 4.0 mm
scatter -0.05 0.1
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 AngioFormer + gradient 0.746 30.2 0.92
2 NeRF-Angio + gradient 0.704 28.4 0.889
3 TV-CS + gradient 0.698 27.99 0.881
4 VesselNet + gradient 0.669 26.61 0.848
5 DiffusionAngio + gradient 0.667 25.83 0.827
6 PnP-ADMM + gradient 0.666 26.04 0.833
7 Learned Primal-Dual + gradient 0.660 25.6 0.821
8 FBPConvNet + gradient 0.653 25.56 0.819
9 FDK + gradient 0.604 24.0 0.769
Spec Ranges (3 parameters)
Parameter Min Max Unit
contrast_timing -0.6 0.9 s
motion -2.4 3.6 mm
scatter -0.06 0.09
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 TV-CS + gradient 0.678 26.94 0.857
2 AngioFormer + gradient 0.677 27.62 0.873
3 NeRF-Angio + gradient 0.665 26.4 0.843
4 PnP-ADMM + gradient 0.639 24.85 0.797
5 Learned Primal-Dual + gradient 0.621 24.65 0.791
6 DiffusionAngio + gradient 0.603 23.98 0.768
7 VesselNet + gradient 0.594 23.62 0.755
8 FBPConvNet + gradient 0.577 22.95 0.729
9 FDK + gradient 0.565 22.11 0.695
Spec Ranges (3 parameters)
Parameter Min Max Unit
contrast_timing -0.35 1.15 s
motion -1.4 4.6 mm
scatter -0.035 0.115
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

About the Imaging Modality

Digital subtraction angiography (DSA) visualizes blood vessels by subtracting a pre-contrast mask image from post-contrast images acquired after injecting iodinated contrast agent. The subtraction eliminates static anatomy, isolating vascular structures. The forward model is y_post - y_pre = Delta_mu * t_vessel + n where Delta_mu is the attenuation increase from iodine. Primary challenges include patient motion between mask and contrast frames, breathing artifacts, and superposition of overlapping vessels.

Principle

X-ray angiography visualizes blood vessels by injecting iodinated contrast agent and acquiring rapid-sequence fluoroscopic images. Digital Subtraction Angiography (DSA) subtracts a pre-contrast mask image from post-contrast frames, removing bone and soft tissue to show only the contrast-filled vasculature with high contrast and spatial resolution.

How to Build the System

Use a biplane or single-plane angiography suite with high-speed flat-panel detectors (30-60 fps capability). The C-arm provides multi-angle positioning. Power injector delivers iodinated contrast (350-370 mgI/mL) at controlled rates. Road-mapping mode overlays vessel map on live fluoro for catheter guidance. 3-D rotational angiography acquires a spin to reconstruct a volume of the vasculature.

Common Reconstruction Algorithms

  • Digital subtraction (mask-live image subtraction)
  • Pixel shifting for motion compensation in DSA
  • 3-D rotational angiography reconstruction (FDK or iterative)
  • Time-density curve analysis for perfusion assessment
  • Deep-learning vessel segmentation and stenosis quantification

Common Mistakes

  • Patient motion between mask and contrast frames causing misregistration artifacts
  • Inadequate contrast bolus timing causing suboptimal vessel opacification
  • Overexposure or underexposure of the detector outside the linear range
  • Bowel gas or cardiac motion causing subtraction artifacts
  • Injecting contrast too fast, creating reflux or missing distal vessels

How to Avoid Mistakes

  • Instruct patients to remain still; use pixel shifting or elastic registration
  • Use test bolus or timing run to determine optimal injection-to-imaging delay
  • Use automatic dose rate control; verify detector within calibrated dynamic range
  • Use cardiac gating for coronary or thoracic angiography
  • Adjust injection rate and volume to vessel size and flow characteristics

Forward-Model Mismatch Cases

  • The widefield fallback applies Gaussian blur, but angiography uses X-ray transmission with iodine contrast agent — the exponential attenuation model with contrast-enhanced vessels is not a simple convolution
  • Digital subtraction angiography (DSA) requires temporal subtraction between pre- and post-contrast images to isolate vessels — the widefield model has no temporal component and cannot model contrast dynamics

How to Correct the Mismatch

  • Use the angiography operator implementing contrast-enhanced X-ray transmission: y = I_0 * exp(-(mu_tissue*t + mu_iodine*c(t))) where c(t) models contrast agent concentration dynamics
  • Apply temporal subtraction (post-contrast minus pre-contrast) or parametric mapping of contrast kinetics using the correct time-resolved forward model

Experimental Setup — Signal Chain

Experimental setup diagram for X-ray Angiography

Experimental Setup

Instrument: Siemens Artis Q / Philips Allura Xper FD20
Image Size: 1024x1024
Kvp: 80
Frame Rate Fps: 30
Contrast Agent: iodinated (Iopamidol 370 mg I/mL)
Injection Rate Ml S: 4.0
Detector: flat-panel (CsI)
Application: cerebral / coronary angiography

Key References

  • Defined by clinical DSA standards (ACC/AHA guidelines)

Canonical Datasets

  • IntrA (intracranial aneurysm 3DRA dataset)

Spec DAG — Forward Model Pipeline

Π(proj) → D(g, η₁)

Π X-ray Projection (proj)
D Flat-Panel Detector (g, η₁)

Mismatch Parameters

Symbol Parameter Description Nominal Perturbed
Δt_c contrast_timing Contrast bolus timing error (s) 0 0.5
σ_m motion Patient motion (mm) 0 2.0
f_s scatter Scatter fraction error 0 0.05

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.