Robust and accurate 2D-tracking-based 3D positioning method: Application to head pose estimation
作者：
Highlights：
• We present a novel, robust and accurate head pose estimation algorithm.
• Our methodology is applicable to any object if we have an approximate 3D model.
• We successfully address the problem of drifting of points that lose track, inherent to most point-tracking systems.
• We validate all the enhancements proposed on two different databases of videos.
• Our final method outperforms the state of the art on the two test databases.
摘要
•We present a novel, robust and accurate head pose estimation algorithm.•Our methodology is applicable to any object if we have an approximate 3D model.•We successfully address the problem of drifting of points that lose track, inherent to most point-tracking systems.•We validate all the enhancements proposed on two different databases of videos.•Our final method outperforms the state of the art on the two test databases.
论文关键词：Pose estimation,Head tracking,Facial point detection and tracking,Outlier correction,POSIT
论文评审过程：Received 8 November 2017, Revised 21 November 2018, Accepted 6 January 2019, Available online 15 January 2019, Version of Record 20 March 2019.
论文官网地址：https://doi.org/10.1016/j.cviu.2019.01.002
