Expression-assisted facial action unit recognition under incomplete AU annotation

作者：

Highlights：

• We recognize AUs with the help of expression labels.

• Expression labels are used as hidden knowledge to complement the missing AU labels.

• The relationships among AUs and expressions are captured by the BN.

• The model is applicable to both complete data and incomplete data.

摘要

Highlights•We recognize AUs with the help of expression labels.•Expression labels are used as hidden knowledge to complement the missing AU labels.•The relationships among AUs and expressions are captured by the BN.•The model is applicable to both complete data and incomplete data.

论文关键词：AU recognition,Incomplete annotation,Bayesian network,Expression

论文评审过程：Received 15 March 2016, Revised 9 July 2016, Accepted 18 July 2016, Available online 21 July 2016, Version of Record 3 August 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.07.028