67,200 individual defects exported in STL format with dimensional values in millimeters. The database was used for CAD model augmentation in X-ray computed tomography simulations for the publication 'Reliability of CT Digital Twins for Defect Segmentation and Probability of Detection Estimation in AM Parts'. It was authored by Catherine Desrosiers and last updated on June 27, 2026.
Use Cases
- Training defect segmentation models based on the large collection of STL-formatted defects.
- Simulating X-ray computed tomography (CT) scans based on CAD-augmented models mentioned in the description.
- Estimating probability of detection for flaws in additive manufacturing parts based on the simulated defect database.
- Benchmarking the reliability of digital twins for non-destructive evaluation based on the described use in the publication.
Strengths
- Contains 67,200 individually exported defects, providing a substantial sample size.
- Each defect includes dimensional values provided in millimeters, offering precise geometric data.
- Specifically created for and documented in a peer-reviewed publication on CT digital twin reliability.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- Borealis Harvested Dataverse
- Collection Method
- Created for CAD model augmentation in X-ray computed tomography simulations.
- Freshness
- Last updated 2026-06-27 04:10:43; freshness should be verified.