1 paper
Hongrui Wu, Zhicheng Gao, Jin Cao +3
Open-vocabulary 3D instance segmentation seeks to segment and classify instances beyond the annotated label space. Existing methods typically map 3D instances to 2D RGB-D images, a…