摘要
目的利用移动终端的传感设备,将广告投放对象精准化。方法采用基于肤色修正的Haar特征人脸检测算法、粒子滤波和Mean-Shift相结合的人脸追踪算法以及Gabor特征性别判别和年龄估计算法,设计基于云计算的硬件架构,展望匿名视频,分析系统发展趋势。结果在云端实现了对广告关注人数、关注时间及关注者年龄、性别的实时统计分析并出具了相关报告。结论对在不同时刻、不同地点投放广告从而获取最大效益的实验提供了有力的数据支撑。
Using sensing devices of mobile terminal,putting the advertisement accuracy will be the most important research direction.Skin fixed Haar feature algorithm for face detection,particle filter with Mean-Shift algorithm for face tracking,Gabor feature for gender discrimination and age prediction are used.It also designs a hardware structure based on Cloud computing,prospects the anonymous video and analyzes the development trend. These make the real-time statistical analysis of the number of advertisement followers,attention time,ages and genders come true.These statistics will provide a powerful basis of when,where and which advertisement to play can get the most effective results.v
出处
《包装工程》
CAS
CSCD
北大核心
2016年第4期129-133,151,共6页
Packaging Engineering
基金
国家科技支撑计划项目(2012BAH85F02)
关键词
广告注意力
人脸检测
性别判别
年龄估计
云计算
advertising attention
face detection
gender discrimination
age prediction
cloud computing
