摘要
多媒体设备上的面部情感分析技术在近期受到了广泛关注,人脸对齐作为一种计算机视觉的面部特征识别技术,受到越来越多研究者和企业的关注和研究。文章对计算机视觉技术中面部情感分析技术的发展和应用进行研究的基础上,深入分析了人脸对齐以及面部情感识别技术。针对面部标定点以及表情之间的关系,设计一种基于DAN网络的面部情感分析模型,采用多步骤的卷积神经网络产生面部标定点,系统分析面部特征点之间的角度以及离散程度来判断面部的情感。分析认为,面部情感分析技术的进一步发展,将被广泛的应用于智慧城市、教育、智能家居、人机交互在内的各个领域,有着较好的应用前景和价值。
Facial sentiment analysis technology on multimedia devices has recently received widespread attention.Face alignment,as a computer vision facial feature recognition technology,has attracted more and more attention and research from researchers and companies.Based on the research on the development and application of facial sentiment analysis technology in computer vision technology,this paper analyzes face alignment and facial sentiment recognition technology in depth.Aiming at the relationship between facial calibration points and expressions,a DAN network-based facial sentiment analysis model is designed.Multi-step convolutional neural networks are used to generate facial calibration points.The system analyzes the angles and the degree of dispersion between facial feature points to determine Facial emotions.According to the analysis,the further development of facial emotion analysis technology will be widely used in various fields such as smart cities,education,smart homes,and human-computer interaction,and has good application prospects and values.
作者
徐天宇
熊超
XU Tian-yu;XIONG Chao(College of Software Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《电脑与信息技术》
2020年第3期13-16,共4页
Computer and Information Technology
关键词
多媒体技术
计算机视觉
深度学习
面部特征点标定
情感分析
multimedia technology
computer vision
deep learning
facial feature point calibration
sentiment analysis
