BEACon: a boundary embedded attentional convolution network for point cloud instance segmentation

Motivated by how humans perceive geometry and color to recognize objects, we propose a boundary embedded attentional convolution (BEACon) network for point cloud instance segmentation. At the core of BEACon, introduce weight in layer adjust neighboring features, with being adapted relationship between changes. As result, BEACon makes use both information, takes as an important feature, thus learns more discriminative feature representation neighborhood. Experimental results show that outperforms state-of-the-art large margin. Ablation studies are also provided prove benefit incorporating into attention