CVPR2021.md

CVPR2021

机器人抓取

ManipulaTHOR: A Framework for Visual Object Manipulation

• Author: Allen Institute for AI / University of Washington
• paper: https://arxiv.org/abs/2104.11213v1

6DoF 位姿估计

FFB6D: A Full Flow Bidirectional Fusion Network for 6D Pose Estimation - oral

这项工作提出了全流双向融合FFB6D网络，根据单个RGBD图像进行6D姿态估计，利用RGB图像中的外观信息和深度图像中的几何信息作为两个互补的数据源，将視覺和几何信息结合，以进行更好的表示学习。在多个数据集上都达到了SOTA！

• Author: Hong Kong University of Science and Technology / Megvii Technology / Kuaishou Technology
• Paper: https://arxiv.org/abs/2103.02242v1
• Code: https://github.com/ethnhe/FFB6D

GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation

• Author: Tsinghua University, BNRist / Technical University of Munich / Google
• Paper: https://arxiv.org/abs/2102.12145v3
• Code: https://github.com/THU-DA-6D-Pose-Group/GDR-Net

FS-Net: Fast Shape-based Network for Category-Level 6D Object Pose Estimation with Decoupled Rotation Mechanism - oral

• Author: , University of Birmingham / Shenzhen University
• Paper: https://arxiv.org/abs/2103.07054v1
• Code: https://github.com/DC1991/FS-Net

Wide-Depth-Range 6D Object Pose Estimation in Space

• Author: EPFL Computer Vision Lab / EPFL Realistic Graphics Lab / ClearSpace SA
• Paper: https://arxiv.org/abs/2104.00337v1

DSC-PoseNet: Learning 6DoF Object Pose Estimation via Dual-scale Consistency

• Author: Baidu Research / ReLER, University of Technology Sydney
• Paper: https://arxiv.org/abs/2104.03658v1

Single-view robot pose and joint angle estimation via render & compare - oral

• Author: ENS Inria / LIGM, ENPC / CIIRC CTU
• HomePage: https://www.di.ens.fr/willow/research/robopose/
• Paper: https://arxiv.org/abs/2104.09359v1