摘要
笔者提出了一种人流智能分析系统,其功能包括人脸检测、人脸性别年龄分析、人脸出现时长分析、人脸表情分析以及人脸总数统计。本研究基于机器学习的人脸开源库,采用多特征提取与卷积神经网络相结合,训练出可以同时进行人脸性别年龄分析、人脸出现时长分析和人脸表情分析的人流智能分析模型,再结合MySQL数据库进行有效统计并生成人流信息数据报表。该技术可被推广到各类经销商铺,来实现对客户特征信息的提取收集及人流动态数据统计。
This paper proposes an artificial intelligence analysis system including face detection, face gender and age analysis, face appearance time analysis, face expression analysis and face total statistics. This research is based on the open source face database of machine learning, and adopts the combination of multi-feature extraction and convolutional neural network to train an artificial intelligence analysis model for face gender and age analysis, face appearance time analysis and facial expression analysis at the same time. In addition, it combines MySQL database to make effective statistics and generate human flow information data report. This technology can be promoted to all kinds of distribution stores to achieve the extraction and collection of customer characteristic information and the flow of dynamic data statistics.
作者
林扬扬
朱文博
陈毓敏
廖家欣
黎毅宇
Lin Yangyang;Zhu Wenbo;Chen Yumin;Liao Jiaxin;Li Yiyu(Department of Automation,Foshan University,Foshan Guangdong 528000,China)
出处
《信息与电脑》
2019年第19期125-128,131,共5页
Information & Computer
基金
国家自然科学基金青年基金项目（项目编号：61804029）
广东省自然科学基金博士启动纵向协同项目（项目编号：2018A030310353）
关键词
人脸识别
人流统计
表情识别
人脸检测
人脸属性
face recognition
flow statistics
facial expression recognition
face detection
face attributes
