SeMoLi: What Moves Together Belongs Together

arXiv:2402.19463v1 Announce Type: new
Abstract: We tackle semi-supervised object detection based on motion cues. Recent results suggest that heuristic-based clustering methods in conjunction with object trackers can be used to pseudo-label instances of moving objects and use these as supervisory signals to train 3D object detectors in Lidar data without manual supervision. We re-think this approach and suggest that both, object detection, as well as motion-inspired pseudo-labeling, can be tackled in a data-driven manner. We leverage recent advances in scene …

3d object abstract arxiv clustering cs.cv data detection instances lidar moving objects results semi-supervised supervision think together train type