Structured dynamic time warping for continuous hand trajectory gesture recognition

作者：

Highlights：

• We use the positions of the corner points in arbitrary trajectories as a kind of structured information, and assign different weights to feature sequences based on the structure.

• We propose a natural continuous trajectory segmentation approach, which combines templates and velocity to detect the beginning and ending points in hand gesture trajectories.

• We combine position and orientation features with DTW to make trajectories more discriminative.

摘要

•We use the positions of the corner points in arbitrary trajectories as a kind of structured information, and assign different weights to feature sequences based on the structure.•We propose a natural continuous trajectory segmentation approach, which combines templates and velocity to detect the beginning and ending points in hand gesture trajectories.•We combine position and orientation features with DTW to make trajectories more discriminative.

论文关键词：Continuous hand gesture recognition,Dynamic time warping,Continuous trajectory segment,Human computer interaction

论文评审过程：Received 3 August 2017, Revised 31 January 2018, Accepted 10 February 2018, Available online 13 February 2018, Version of Record 7 March 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.02.011