Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

X-Gaussian logo

Radiative Gaussian Splatting for Efficient X-ray Novel View Synthesis

ECCV 2024

arXiv 知乎 Project Page AK GitHub HF Model HF Demo

5 CT scenes + 3 real scans · X-ray novel view synthesis · Direct 3D CT reconstruction

Demo · Paper · Project Page · Code · Models · Data

Chest CT reconstruction during optimization Foot CT reconstruction during optimization Head CT reconstruction during optimization Abdomen CT reconstruction during optimization
Pancreas CT reconstruction during optimization Pine CT reconstruction during optimization Seashell CT reconstruction during optimization Walnut CT reconstruction during optimization

CT Reconstruction During Optimization

Interactive Demo

Explore eight pretrained scenes or upload an X-Gaussian PLY in our Gradio demo. Rotate the Gaussian point cloud, move its linked CT slice, render X-ray projections, and export an interactive mesh.

X-Gaussian interactive demo

For local installation and usage, see the code repository.

Dataset Description

This repository hosts the five processed CT scenes used in X-Gaussian, introduced in Radiative Gaussian Splatting for Efficient X-ray Novel View Synthesis (ECCV 2024): Chest, Abdomen, Foot, Head, and Pancreas.

Each scene contains a reference 3D CT volume, simulated cone-beam X-ray projections, view angles, and scanner geometry. The data support sparse-view reconstruction and held-out X-ray projection rendering. The current X-Gaussian code also supports direct 3D CT reconstruction from the same Gaussian field.

Exported Meshes - Click each to open its interactive viewer

Chest exported mesh Foot exported mesh Head exported mesh Abdomen exported mesh
Pancreas exported mesh Pine exported mesh Seashell exported mesh Walnut exported mesh

Performance

X-ray Novel View Synthesis

Original paper NVS comparison

Original paper NVS visual results

Direct 3D CT Reconstruction

The CT extension uses the original Gaussian point attributes and native data resolution, with 50 training projections per scene. Synthetic recipes use 30,000 base steps and 15,000 refinement steps; real recipes use 30,000 joint projection/3D-TV steps. Settings were selected on held-out training projections. Reference volumes and test projections are used only for evaluation.

3D PSNR uses a fixed range of 1 without clipping or intensity correction. SSIM averages nonempty reference slices along each axis, then across axes. Synthetic references are the original CT volumes; real references are full-view FDK pseudo-volumes, not measured CT ground truth.

Scene CT resolution 3D PSNR (dB) 3D SSIM
Chest 128 × 128 × 128 36.00 0.9391
Abdomen 512 × 512 × 463 34.27 0.8946
Foot 256 × 256 × 256 32.18 0.8815
Head 256 × 256 × 128 41.44 0.9768
Pancreas 512 × 512 × 240 22.38 0.8791
Average 33.25 0.9142

Real Data 3D CT Reconstruction

Real FIPS scans contain 50 training and 100 test projections at 560 × 560. The 3D reference is full-view FDK.

Scene CT resolution 3D PSNR (dB) 3D SSIM
Pine 256 × 256 × 256 41.80 0.9465
Seashell 256 × 256 × 256 44.01 0.9638
Walnut 256 × 256 × 256 32.77 0.7169
Average 39.53 0.8757

Real CT reconstruction

See the real-data sources and attribution.

Dataset Contents

|--data
      |--synthetic
            # The first five datasets are used in our paper
            |--chest_50.pickle
            |--abdomen_50.pickle
            |--foot_50.pickle
            |--head_50.pickle
            |--pancreas_50.pickle
      |--real
            |--pine
            |--seashell
            |--walnut

Training-only FDK initialization is also included as synthetic/init_<scene>_50.npy and real/<scene>/init_<scene>.npy. Training reuses these files when present and computes initialization automatically when absent.

One Python pickle file is provided per original CT scene under synthetic/. The real scans use the folder format described above. Dimensions below are the original stored array dimensions; no downsampling is applied to this release.

File CT volume shape Projection shape per view
synthetic/chest_50.pickle 128 × 128 × 128 256 × 256
synthetic/abdomen_50.pickle 512 × 512 × 463 1024 × 1024
synthetic/foot_50.pickle 256 × 256 × 256 512 × 512
synthetic/head_50.pickle 256 × 256 × 128 512 × 512
synthetic/pancreas_50.pickle 512 × 512 × 240 1024 × 1024
real/pine/ 256 × 256 × 256 560 × 560
real/seashell/ 256 × 256 × 256 560 × 560
real/walnut/ 256 × 256 × 256 560 × 560

Each synthetic pickle contains 50 training projections and 50 held-out projections. The held-out split is named val; these are different views of the same scene, rather than separate training and validation scenes. The synthetic files specify cone-beam geometry, a configured scan range of 180°, and noise: 0.

Real scenes contain 50 training and 100 held-out measured projections; their listed volumes are full-view FDK pseudo-references.

Synthetic Data Format

Field Description
image Reference CT volume with shape (N_x, N_y, N_z), used for reconstruction evaluation.
train.projections, val.projections Projection arrays with shape (50, H, W).
train.angles, val.angles Corresponding view angles in radians.
numTrain, numVal Numbers of training and held-out views.
DSD, DSO Source-to-detector and source-to-origin distances.
nDetector, dDetector Detector pixel counts and pixel sizes.
nVoxel, dVoxel CT voxel counts and voxel sizes.
offOrigin, offDetector Volume and detector offsets.
mode, accuracy, filter Acquisition and reconstruction settings.

Physical lengths in the stored scanner metadata are in millimeters. The official loader handles unit conversion and scene normalization. CT ground truth is not required by the current X-Gaussian training workflow.

Usage

Please refer to the X-Gaussian repository for installation, data preparation, training, and testing.

Data Sources and Processing

The original five CT scenes use the processed CT data format shared with SAX-NeRF. X-ray projections are synthesized from CT volumes using the TIGRE-based data generation pipeline. See the SAX-NeRF dataset card and data generation code for source attribution and preprocessing details. Acquisition settings for this release should be read from the individual files as summarized above.

Citation

If you use these data, please cite X-Gaussian and SAX-NeRF:

@inproceedings{x_gaussian,
  title={Radiative gaussian splatting for efficient x-ray novel view synthesis},
  author={Yuanhao Cai and Yixun Liang and Jiahao Wang and Angtian Wang and Yulun Zhang and Xiaokang Yang and Zongwei Zhou and Alan Yuille},
  booktitle={ECCV},
  year={2024}
}

@inproceedings{sax_nerf,
  title={Structure-Aware Sparse-View X-ray 3D Reconstruction},
  author={Yuanhao Cai and Jiahao Wang and Alan Yuille and Zongwei Zhou and Angtian Wang},
  booktitle={CVPR},
  year={2024}
}

License

The original five CT datasets retain the Apache-2.0 license. The FIPS real data and derived initialization under real/ are CC BY 4.0; see source attribution and processing details.

Downloads last month
480

Models trained or fine-tuned on CaiYuanhao/X-Gaussian

Space using CaiYuanhao/X-Gaussian 1

Paper for CaiYuanhao/X-Gaussian