All datasets
AvailableXRAY

Veriserum

Veriserum is a dual-plane fluoroscopic X-ray dataset of knee implant phantoms designed for two-dimensional to three-dimensional registration, implant pose estimation, segmentation and reconstruction research.

Licence and access

ETH Zurich describes Veriserum as freely accessible and open source for academic and research use. The official project page does not state a general Creative Commons license permitting unrestricted commercial reuse, so commercial use is treated as restricted pending permission from the dataset owners.

open accesscommercial use restrictedredistribution allowed
Visit official sourcelast verified 2026-09-14
Coverage
kneeno labelsphantom, not human anatomy
subjects
not verified
images
110,000
formats
not verified
origin
not verified
Citation
Wang J, Vogl F, Schütz P, Ćuković S, Taylor WR. Veriserum: A dual-plane fluoroscopic dataset with knee implant phantoms for deep learning in medical imaging. arXiv. 2025. DOI: 10.48550/arXiv.2509.05483. Dataset DOI: 10.3929/ethz-b-000701146.
Notes

The official ETH Zurich page reports approximately 110,000 X-ray images acquired during 1,600 trials using ten femoral-tibial implant combinations. All images have automatically registered pose information, and 200 images additionally have manually registered reference poses. Because the images show implant phantoms rather than human participants, num_subjects is null. Pose annotations do not fit the current classification-or-segmentation label enum, so labels is recorded as none. The exact image-file extension was not stated clearly enough to populate the format field.

This record is a pointer. Licence and access terms are recorded as they stood on the verification date above; what binds you is the dataset's own licence text. Confirm at the source before use.

Orthopedic AI Research Updates

Monthly research digest, product updates, and clinical AI insights.

Unsubscribe anytime.

Veriserum, dataset record