Face expression identification method and device

• Main Authors: LUO QINGGANG, XU LILI, LI YANGYANG, GAO QIAN, CHEN ZENGZHAO, YANG FAN, HE XIULING, FANG JING, JIANG LANG, WANG YAOYAN, WU KE, ZHU SHUPEI, ZHOU MENGYING
• Format: Patent
• Language: Chinese; English
• Published: 06.07.2018
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention provides a face expression identification method and device and relates to the face expression identification technology field. The method comprises steps that a peak frame image of a to-be-identified face expression image sequence is acquired; geometric characteristics of the peak frame image are extracted; Gabor characteristics of the peak frame image are extracted; based on the geometric characteristics and the Gabor characteristics, integrated characteristics are acquired; the integrated characteristics are taken as input, a preset depth multi-kernel learning model is trained, and a kernel function outputted by the multi-kernel learning model after training is acquired; the kernel function is inputted to a preset classifier for expression classification, and the identification result is acquired. The method is advantaged in that the identification rate is improved, and efficiency is higher. 本发明实施例提供了种脸部表情识别方法及装置，涉及脸部表情识别技术领域。所述方法包括获取待识别脸部表情图像序列的峰值帧图像；提取所述峰值帧图像的几何特征；提取所述峰值帧图像的Gabor特征；基于所述几何特征