PeP: a Point enhanced Painting method for unified point cloud tasks

arXiv:2310.07591v2 Announce Type: replace
Abstract: Point encoder is of vital importance for point cloud recognition. As the very beginning step of whole model pipeline, adding features from diverse sources and providing stronger feature encoding mechanism would provide better input for downstream modules. In our work, we proposed a novel PeP module to tackle above issue. PeP contains two main parts, a refined point painting method and a LM-based point encoder. Experiments results on the nuScenes and KITTI datasets validate the …

abstract arxiv cloud cs.cv diverse encoder encoding feature features importance modules novel painting pipeline recognition tasks type vital work