首页|Reports on Robotics Findings from Shanghai Jiao Tong University Provide New Insi ghts (Attentionvote: a Coarse-to-fine Voting Network of Anchor-free 6d Pose Esti mation On Point Cloud for Robotic Bin-picking Application)
Reports on Robotics Findings from Shanghai Jiao Tong University Provide New Insi ghts (Attentionvote: a Coarse-to-fine Voting Network of Anchor-free 6d Pose Esti mation On Point Cloud for Robotic Bin-picking Application)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Current study results on Robotics have been published. According to news reporting originating from Shanghai,People's Republic of China,by NewsRx correspondents,research stated,"Current state-of -the-art pose estimation methods are almost launched on segmented RGB-D images. However,these methods may not apply to more general industrial parts due to a l ack of texture information and highocclusion of stacked objects." Financial support for this research came from National Natural Science Foundatio n of China (NSFC). Our news editors obtained a quote from the research from Shanghai Jiao Tong Univ ersity,"This article establishes an end-to-end pipeline to synchronously regres s all potential object poses from an unsegmented point cloud. The point pair fea tures (PPFs) are first extracted and then fed into a PointNet-like backbone for obtaining the point-wise features. Based on the center voting,a coarse-to-fine voting architecture is proposed to extract instance features instead of implemen ting instance segmentation. A lightweight threedimensional (3D) heatmap is leve raged to cluster votes and generate center seeds. Further,an attention voting m odule is constructed to fuse point-wise features into instance-wise features ada ptively. Ultimately,the suggested network regresses object poses with a quatern ion loss to handle the symmetric puzzle. The network holds the advantage of prod ucing the final pose prediction without any post-processing steps like nonmaximu m suppression (NMS) or any pose refinement modules like iterative closest point (ICP). The proposed network is evaluated on the public Fraunhofer IPA dataset,w hich demonstrates the robustness of the pose estimation network with much better performance."
ShanghaiPeople's Republic of ChinaAs iaEmerging TechnologiesMachine LearningRoboticsRobotsShanghai Jiao Tong University
NETL
NSTL
