摘要
论文探讨了一种新的多通道统计分析方法—独立分量分析方法(Independent Component Analysis,ICA)。简要介绍了独立分量分析问题的描述和预处理,分析了基于负熵最大化判据的目标函数和FastICA算法,并将其应用于人脸特征提取,在人脸识别中得到了较为满意的结果。
Independent Component Analysis(ICA)of a new multidimensional signal processing method is discussed.In this pa⁃per,the description of problem and pretreatment in ICA is introduced briefly,the object function and FastICA algorithm based on negentropy maximization criterion are analyzed,and it is applied to face feature extraction,satisfactory result is gotten in face recog⁃nization.
作者
赵民全
ZHAO Minquan(No.92785 Troops of PLA,Huludao 125208)
出处
《舰船电子工程》
2021年第2期36-38,68,共4页
Ship Electronic Engineering
关键词
独立分量分析
负熵
分离矩阵
特征提取
independent component analysis
negentropy
separating matrix
feature extraction
