Occlusion-robust method for RGB-D 6-DOF object tracking with particle swarm optimization

作者：

Highlights：

• Particle swarm optimization is used to track 3D objects with six degrees of freedom.

• Tracking is markerless, model-based, and uses generic features.

• Experiments using a database of real environments with challenging image sequences.

• The proposed technique with excellent accuracy and robustness to occlusion.

• Method compared with state-of-the-art optimization-based and learning-based trackers.

摘要

•Particle swarm optimization is used to track 3D objects with six degrees of freedom.•Tracking is markerless, model-based, and uses generic features.•Experiments using a database of real environments with challenging image sequences.•The proposed technique with excellent accuracy and robustness to occlusion.•Method compared with state-of-the-art optimization-based and learning-based trackers.

论文关键词：3D object tracking,Particle swarm optimization,RGB-D images

论文评审过程：Received 15 December 2020, Revised 30 January 2021, Accepted 13 February 2021, Available online 23 February 2021, Version of Record 6 March 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114736