基于人脸特征定位的SNS网站应用组件研究与设计
2017-01-12吕鑫李月军
吕鑫+李月军



摘 要: 针对国内SNS网站进行调研,分析用户群体特征及社会心理,剖析SNS网站流行的原因,在此基础上提出了在SNS网站上运营好一款人气高、粘性强的应用组件。实验研究并分析了主动形状模型(ASM)和受限局部模型(CLM)的优缺点,提出将两者优点结合的人脸面部特征定位提取方法,设计开发了一款应用组件产品。通过用户上传照片或在线拍摄,捕捉出人脸头像及面部特征,并按照用户在图像库中选择的照片进行合成,生成一张具有两张人脸特征的新头像。将组件接入SNS网站,通过一段时间的运营数据分析,结合互联网的实际应用,提出对组件进行优化的方案。
关键词: 人脸检测; 特征定位; SNS网站; 应用组件
中图分类号: TN911.7?34; TM417 文献标识码: A 文章编号: 1004?373X(2016)23?0170?04
Research and design of SNS website application component based on face feature location
L? Xin, LI Yuejun
(College of Optical and Electronical Information Changchun University of Science and Technology, Changchun 130000, China)
Abstract: The SNS website in China is investigated to analyze the user group characteristics and social psychology, and dissect the popular reasons for the SNS website. On this basis, how to better operate a high?popularity and strong?stickiness application component on SNS website is proposed. The advantages and disadvantages of the active shape model (ASM) and CLM are studied and analyzed by the experiments. The face feature location extraction method combining with the advantages of the models is proposed. An application component product was designed and developed. The head portrait and face feature are captured by means of the user photos upload and online shoot. The photos selected in the image library by users are synthesized to generate a new head portrait with two face features. The component is accessed into the SNS website. On the basis of analyzing the a period of operation data and combining with the practical application of the Internet, the optimization scheme of the component is proposed.
Keywords: face detection; feature location; SNS website; application component
0 引 言
人脸识别是当前人工智能和模式识别的研究热点,就其技术本质而言,是通过图像采集设备获取用户的面部图像,再对其脸部的特征进行计算分析,进而和自身数据库里已有的样本进行比对,最后判断出用户的真实身份。人脸识别是当前人工智能和模式识别的研究热点,在家庭娱乐、互联网等领域已经有了一些有趣的应用,比如有一些智能玩具、机器人能够识别家人的身份;某些游戏具备玩家真实面像等[1]。但目前人脸识别的广泛应用性较弱,尤其是针对当今流行的SNS社交网络平台。
1 基于SNS网站用户的需求调研
1.1 用户基本特征分析
目前国内主要的SNS网站用户中,男性用户数量显著高于女性用户,其在男性网民中的渗透率也高于女性网民。……
