文档介绍:
人脸特征点定位在微表情提取中的应用#
郭珈瑜1,舒茂1,鲁荣荣1,张瑒1,杨卓1,代煜2**
(1. 南开大学信息技术科学学院,天津 300071;
5
10
15
20
25
30
2. 南开大学机器人与信息自动化研究所,天津 300071)
摘要:微表情是一种不易被人察觉,转瞬即逝的表情。它往往是在人刻意压制自己内心情感
流露的时候产生的一种痛苦的表情模式,不受人的意识控制。本文旨在介绍一种基于计算机
视觉处理技术,通过准确定位人脸特征点来检测并提取视频流中的微表情图片的方法,该方
法中人脸特征点定位准确,可靠,微表情识别具有较高的准确率。
关键词:计算机视觉;微表情;人脸特征点定位;最小邻域均值投影
中图分类号:TP391
Applications of human face features localization in
extraction for micro-expression
GUO Jiayu1, SHU Mao1, LU Rongrong1, ZHANG Yang1, YANG Zhuo1, DAI Yu2
(1. College of Information Technical Science, Nankai University, TianJin 300071;
2. Robot and Information Automation Institute of Nankai University, TianJin 300071)
Abstract: Micro-expression is not easy to be aware of and it lasts for a short time. It is often a kind
of painful expression pattern which is produced when people deliberately suppress their feelings.
The purpose of this paper is to introduce a method, based puter vision processing
technology and with face features accurate localization , to detect and extract micro-expression
images in a video streaming, In the method , localization of face features is accurate and reliable,
Micro-expression recognition has high accuracy.
Keywords: Computer vision; Micro-expression; Localization of facial feature points; minimal
neighborhood mean projection function
0 引言
表情作为人类表达自身情感信息的重要载体,可视为人类心理活动的晴雨计。人类对
于面部表情的研究已经非常丰富,Haggard 和