SeMoLi is a data-driven method for pseudo-labeling moving objects in Lidar data, outperforming heuristic-based approaches and demonstrating cross-dataset generalization.
