Self-supervised deep learning models can accurately perform 3D segmentation of cell nuclei in complex biological tissues, enabling scalable analysis in settings with limited or no ground truth ...