基于人脸识别技术的智慧营业厅系统研究与应用  

Research and Application of Intelligent Business Hall System Based on Face Recognition Technology

作　　者：刘慧娟 赵东明[2] LIU Hui-juan;ZHAO Dong-ming(Tianjin Broadcasting TV and Film Institute,Tianjin 300112;China Mobile Communications Group Tianjin Co.,Ltd.,Tianjin 300020)

机构地区：[1]天津广播影视职业学院影视技术系,天津300112 [2]中国移动通信集团天津有限公司,天津300020

出　　处：《数字技术与应用》2019年第2期89-90,共2页Digital Technology & Application

摘　　要：本文提出一种面向智慧营业厅的客户人脸识别技术方法,基于高清摄像头和人工智能视频处理程序,结合机器视觉视频处理框架实现客户到厅的身份自动识别,利用CNN深度神经网络不断优化识别模型,通过比对进厅用户与当日基站用户的人脸特征向量集合,快速识别用户身份,给予针对性的画像、标签,轻渠道精准营销推荐和潜在服务意图挖掘,增强了客户感知。客户人脸识别模型完全以亚洲人脸特征为基础,以移动用户图像为数据源,模型准确率高且适配性强,显著提升了营业效率。This paper presents a method of customer face recognition technology for intelligent business hall.Based on high-definition camera and artificial intelligence video processing program,combined with Machine Vision video processing framework,automatic identification of customer arriving at the hall is realized.Convolutional neural network is used to optimize the recognition model continuously.By comparing the set of face feature vectors between the user entering the hall and the user of the base station on the same day,Fast identification of user identity,targeted portraits and labels,light channels of accurate marketing recommendation and potential service intention mining,enhance customer perception.The customer face recognition model is based entirely on Asian face features and uses CMCC user images as data sources.The model has high accuracy and adaptability,which significantly improves business efficiency.

关 键 词：人脸识别 智慧营业厅 卷积神经网络 精准推荐 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]