摘要
采用主成分分析(PCA)作为统计分析方法的主动表现模型(AAM)是建立二维可形变模型的有效方法。提出一种将改进的AAM用于人脸面部特征定位的新方法,并与传统AAM进行比较,实验证明此方法要优于传统AAM。
Active Appearance Model （AAM） is an effective method to build 2D deformable model for an object, which adopts Principal Component Analysis （PCA） as its statistical analysis. In this paper, a new approach of facial feature location based on improved AAM is proposed and compared with the facial feature location based on standard AAM. The experimental results demonstrate that this method has better performance than standard AAM.
出处
《电视技术》
北大核心
2009年第7期84-86,共3页
Video Engineering
基金
教育部博士点基金项目(20060359004)
教育部留学归国人员科研启动基金(413117)
关键词
独立成份分析
主动表现模型
主成份分析
人脸面部特征定位
independent component analysis
active appearance model
principal component analysis
facial feature location
