SkullBreak - Dataset for Automatic Cranial Implant Design

The SkullBreak dataset is intended for the development of data-driven solutions for cranial implant design, which remains to be a time-consuming and laborious task in current clinical routine of cranioplasty.

It was first used in the MICCAI 2021 AutoImplant Challenge (https://autoimplant2021.grand-challenge.org/) together with the SkullFix dataset (https://doi.org/10.1016/j.dib.2021.106902).

The dataset was adapted from the public head CT collection CQ500 with CC BY-NC-SA 4.0 license and it contains 114 and 20 complete skulls, each accompanied by defective skulls and the corresponding cranial implants, for training and evaluation respectively. On each skull, five different synthetic defects were created:

• unilateral defect in the parieto-temporal area,
• unilateral defect in the fronto-orbital area,
• bilateral defect,
• two random defects.

This resulted in a training set with 570 triplets (complete skull, defective skull and the implant) and a test set with 100 triplets. The defects in this track were created with random shapes.

A reference to the following publication is required to use any part of the SkullBreak dataset:

• Kodym, O., Li, J., Pepe, A., Gsaxner, C., Chilamkurthy, S., Egger, J. and Španěl, M., 2021. SkullBreak/SkullFix–Dataset for automatic cranial implant design and a benchmark for volumetric shape learning tasks. Data in Brief, 35, p.106902. (https://doi.org/10.1016/j.dib.2021.106902).