People re-identification using skeleton standard posture and color descriptors from RGB-D data
作者：
Highlights：
• A people re-identification approach by color and depth data is proposed.
• Point clouds are aligned to generate pose-independent signatures.
• An unevenly-spaced partition grid is computed on the skeleton standard posture of the person under analysis. The partition grid drives a color-based descriptor extraction.
• The signatures of reference database are re-projected using the partition grid of the person under analysis to highlight shape differences.
摘要
•A people re-identification approach by color and depth data is proposed.•Point clouds are aligned to generate pose-independent signatures.•An unevenly-spaced partition grid is computed on the skeleton standard posture of the person under analysis. The partition grid drives a color-based descriptor extraction.•The signatures of reference database are re-projected using the partition grid of the person under analysis to highlight shape differences.
论文关键词：People re-identification,Color-based descriptor,Skeleton standard posture,Partition grid,RGB-D sensor,Color point cloud
论文评审过程：Received 17 October 2017, Revised 21 November 2018, Accepted 4 January 2019, Available online 5 January 2019, Version of Record 8 January 2019.
论文官网地址：https://doi.org/10.1016/j.patcog.2019.01.003
