Improving Synthetic Data Generation Through Federated Learning in Scarce and Heterogeneous Data Scenarios | Publicación