DSRF: A flexible trajectory descriptor for articulated human action recognition

作者：

Highlights：

• We propose a novel articulated skeleton representation by modeling the skeleton as the interconnections of multiple rigid bodies.

• 6-D rigid body motion trajectories are represented by the proposed invariant Dual Square-Root Function (DSRF) descriptor.

• The Virtual Rigid Body (VRB) configuration is introduced to produce compact mid-level features for representing the movement of each part.

• We develop a Most Informative Part (MIP) method to select a subset of consistency and activity body parts in the final skeletal representation.

摘要

•We propose a novel articulated skeleton representation by modeling the skeleton as the interconnections of multiple rigid bodies.•6-D rigid body motion trajectories are represented by the proposed invariant Dual Square-Root Function (DSRF) descriptor.•The Virtual Rigid Body (VRB) configuration is introduced to produce compact mid-level features for representing the movement of each part.•We develop a Most Informative Part (MIP) method to select a subset of consistency and activity body parts in the final skeletal representation.

论文关键词：Articulated human action recognition,Rigid body motion trajectory,Dual square root function descriptor,Most informative part,Virtual rigid body

论文评审过程：Received 13 December 2016, Revised 11 October 2017, Accepted 24 October 2017, Available online 26 October 2017, Version of Record 6 November 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.10.034