Table 10. Performance comparison based on BraTS dataset.

No.	Author	Pretext task	Category	Random init.: DSC (%)	SSL: DSC (%)
1	Zhou et al. (2019)	Models Genesis	Generative	90.68	92.58
2	Taleb et al. (2021)	Jigsaw Puzzle	Predictive	80.54	89.74
3	Zhu et al. (2020a)	Rubik Cube+	Predictive	85.47	89.6
4	Chen et al. (2019)	Context Restoration	Generative	84.41	85.57
5	Zhang et al. (2020)	Scale Aware Rest.	Multi-Tasking	74.35	84.92
6	Chen et al. (2019)	Image Inpainting*	Generative	84.41	84.54
7	Taleb et al. (2020)	3D Relative Pos. Pred.	Predictive	76.38	81.28
8	Taleb et al. (2020)	3D CPC	Contrastive	76.38	80.83
9	Taleb et al. (2020)	3D Rotation	Predictive	76.38	80.21
10	Taleb et al. (2020)	3D Jigsaw	Predictive	76.38	79.66
11	Taleb et al. (2020)	3D Exemplar	Predictive	76.38	79.46