A hierarchical model for learning to understand head gesture videos

作者：

Highlights：

• Propose a hierarchical model to understand head gesture videos.

• Utilize the multi-task learning framework.

• Compress features using stacked BLSTM.

• Propose several applications in multiple fields.

摘要

•Propose a hierarchical model to understand head gesture videos.•Utilize the multi-task learning framework.•Compress features using stacked BLSTM.•Propose several applications in multiple fields.

论文关键词：Video understanding,Head gesture videos,Stacked BLSTM,Multi-task learning,Transfer learning

论文评审过程：Received 4 September 2019, Revised 8 April 2021, Accepted 12 August 2021, Available online 14 August 2021, Version of Record 19 August 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108256