Label-Efficient Self-Supervised Federated Learning for Tackling Data Heterogeneity in Medical Imaging | IEEE Journals & Magazine | IEEE Xplore