Table 5. Summary of contrastive self-supervised learning methods in medical imaging.
No. | Authors | Pretext task | Down-stream task |
---|---|---|---|
1 | Jamaludin, Kadir & Zisserman (2017) | Longitudinal spinal MRI | Disc degeneration grading |
2 | Lu, Chen & Mahmood (2020) | CPC | Breast cancer classification |
3 | Zhu et al. (2020b) | TCPC | Brain hemorrhage classification Lung Nodule classification |
4 | Xie et al. (2020) | BYOL | Liver segmentation Spleen segmentation Kidney tumour seg. Abdominal organs seg. |
5 | Li et al. (2020a) | Feature-based softmax embedding | PM classification AMD classification Diabetic retinopathy detection |
6 | Sowrirajan et al. (2021) | MoCo | Tuberculosis detection Pleural effusion classification |
7 | Vu et al. (2021) | MoCo | Pleural effusion classification |
8 | Sriram et al. (2021) | MoCo | COVID patient prognosis |
9 | Chen et al. (2021b) | MoCo | COVID few-shot classification |
10 | Chaitanya et al. (2020) | SimCLR | Cardiac segmentation Prostate segmentation |
11 | Azizi et al. (2021) | SimCLR | Chest X-ray classification Skin lesions classification |