Cascade of Tasks for facial expression analysis
作者：
Highlights：
• Proposed a method combining three complementary detection tasks to detect AU events rather than frames.
• Performed best in four public datasets that differ in complexity.
• Proposed a novel event-based evaluation metric.
• Experimental results are reported in both conventional frame-based and new event-based metrics.
• Proposed an efficient search approach to better suit long image sequences.
摘要
•Proposed a method combining three complementary detection tasks to detect AU events rather than frames.•Performed best in four public datasets that differ in complexity.•Proposed a novel event-based evaluation metric.•Experimental results are reported in both conventional frame-based and new event-based metrics.•Proposed an efficient search approach to better suit long image sequences.
论文关键词：Automated facial expression analysis,Action unit detection,FACS
论文评审过程：Received 21 February 2014, Revised 28 August 2015, Accepted 22 March 2016, Available online 31 March 2016, Version of Record 7 May 2016.
论文官网地址：https://doi.org/10.1016/j.imavis.2016.03.008
