A spatial-temporal framework based on histogram of gradients and optical flow for facial expression recognition in video sequences
作者：
Highlights：
• Proposes a facial expression recognition framework that uses dynamic information.
• Dynamic information is variation in facial shape and movement of facial landmarks.
• Introduces PHOG_TOP to capture variation of facial shape in temporal domain.
• Introduces dense optical flow on a grid to detect movement of landmarks.
• Fusion of PHOG_TOP and dense optical flow for high recognition rate.
摘要
Highlights•Proposes a facial expression recognition framework that uses dynamic information.•Dynamic information is variation in facial shape and movement of facial landmarks.•Introduces PHOG_TOP to capture variation of facial shape in temporal domain.•Introduces dense optical flow on a grid to detect movement of landmarks.•Fusion of PHOG_TOP and dense optical flow for high recognition rate.
论文关键词：Histogram of gradients,Facial expression,Optical flow,Feature extraction
论文评审过程：Received 13 October 2014, Revised 31 March 2015, Accepted 27 April 2015, Available online 8 May 2015, Version of Record 16 July 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.04.025
