Self-supervised learning for Segmentation This work proposes a self-supervised learning framework for kidney segmentation, leveraging anotomical asymmetry of kidneys as a novel proxy task. Please see our paper Self-Supervised Learning for Segmentation for details.