Subject	Information
Specific subject area	Computer Vision and Pattern Recognition
Type of data	Image
How data were acquired	The skulls are segmented from head computed tomography (CT) scans using a customized thresholding technique.
Data format	Raw
Parameters for data collection	The selection of DICOM files was based on the image quality (e.g., slice thickness, fracture, scanning protocol).
Description of data collection	The dataset includes two types of skulls: the 500 healthy skulls, each of which contains the complete bony structures of a human skull and the 29 craniotomy skulls, where a part of the cranial bone is missing on each skull.
Data source location	Medical University of Graz
Data accessibility	The download link of this dataset can be found from the Figshare repository1: https://figshare.com/s/e3d9debd55ad24c84678?file=17264471
Related research articles	jianning Li, Gord von Campe, Antonio Pepe, Christina Gsaxner, Enpeng Wang, Xiaojun Chen, UlrikeZefferer, Martin Tödtling, Marcell Krall, Hannes Deutschmann, et al. Automatic skull defect restoration andcranial implant generation for cranioplasty. Medical Image Analysis, 73:102171, 2021. DOI: https://doi.org/10.1016/j.media.2021.102171. reference: [1]