4D-CS On the multi-scan semantic segmentation benchmark of the semanticKITTI dataset, our method ranks 1st among methods that only use LiDAR points. We also achieve SOTA result on moving objects segmentation on SemanticKITTI dataset.