FaCT-GS: Volumetric CT Scans and Synthetic Projections for Reconstruction

A 22.2 GB collection of volumetric CT scans and synthetic projections compiled for the FaCT-GS research project. The dataset aggregates data from multiple public repositories including LIDC-IDRI, SciVis, Pancreas-CT, X-Plant, and FIPS, and was last updated on 2026-04-24. It was shared by author Pawel Tomasz Pieta in conjunction with the associated research paper.

Use Cases

• Training and evaluating novel CT reconstruction models based on the provided volumes and projections.
• Benchmarking 3D Gaussian splatting methods for medical and scientific visualization using the synthetic and real scan data.
• Developing domain adaptation techniques for CT reconstruction across different anatomical structures like chest, head, and plants.

Strengths

• Dataset size is 22.2 GB, indicating a substantial volume of data for model training.
• Explicitly lists original data sources and licenses, providing clear provenance.
• Includes both synthetic projections and real CT scans, offering varied data types for testing.

Limitations

• Specific column names and row counts are unknown, limiting pre-download assessment of data structure.
• Data is aggregated from multiple sources with different licenses, requiring careful compliance checks.

Provenance

Source

Aggregated from public repositories: LIDC-IDRI, SciVis, Pancreas-CT, X-Plant, FIPS, and a private coral scan.

Collection Method

Combined from existing public datasets and supplemented with synthetic projections.

Freshness

Last updated 2026-04-24 11:51:13.