FPCC: Fast point cloud clustering-based instance segmentation for industrial bin-picking

Instance segmentation is an important pre-processing task in numerous real-world applications, such as robotics, autonomous vehicles, and human–computer interaction. Compared with the rapid development of deep learning for two-dimensional (2D) image tasks, learning-based instance 3D point cloud still has a lot room development. In particular, distinguishing large number occluded objects same class highly challenging problem, which seen robotic bin-picking. usual bin-picking scene, many identical are stacked together model known. Thus, semantic information can be ignored; instead, focus put on instances. Based this requirement, we propose Fast Point Cloud Clustering (FPCC) industrial scene. FPCC includes network named FPCC-Net fast clustering algorithm. extracts features each infers geometric center points simultaneously. After that, proposed algorithm clusters remaining to closest feature embedding space. Experiments show that also surpasses existing works scenes more computationally efficient. Our code data available at (https://github.com/xyjbaal/FPCC).