Labeling Vertebrae with Two-dimensional Reformations of Multidetector CT Images: An Adversarial Approach for Incorporating Prior Knowledge of Spine Anatomy

Anjany Sekuboyina, Markus Rempfler, Alexander Valentinitsch, Bjoern Menze, Jan S. KirschkeRadiology 2020 · DOI

Abstract

An identification performance comparable to existing three-dimensional approaches was achieved when labeling vertebrae on two-dimensional MIPs. The performance was further improved using the proposed adversarial training regimen that effectively enforced local spine a priori knowledge during training. Spine localization increased the generalizability of our approach by homogenizing the content in the MIPs.<i>Supplemental material is available for this article.</i>© RSNA, 2020.
BibTeX
@inproceedings{sekuboyina2020labeling,
  author = {Anjany Sekuboyina and Markus Rempfler and Alexander Valentinitsch and Bjoern Menze and Jan S. Kirschke},
  title = {{Labeling Vertebrae with Two-dimensional Reformations of Multidetector CT Images: An Adversarial Approach for Incorporating Prior Knowledge of Spine Anatomy}},
  booktitle = {Radiology},
  year = {2020},
  doi = {10.1148/ryai.2020190074},
}

Loading related content...

Figure