NeurIPS 2020

3D Self-Supervised Methods for Medical Imaging

Meta Review

Applying self-supervision methods to the medical domain is interesting to the community, and adaptation of the methods from 2D to 3D domain is also of value in the machine learning community. There are numerous self-supervision works on medical images mentioned by R1 post rebuttal. The authors are encouraged to include those in the paper, and properly place their paper in the context of previous self-supervision work in the medical imaging domain.