摘要
步态识别是生物特征识别技术中的一个新兴领域,它根据人们走路的个体特点进行身份识别,具有非侵犯性、难以隐藏、对系统分辨率要求低、远距离识别等优点,已成为基于视觉的人体运动分析的研究热点。该文提出了基于主成分分析(PCA)的特征提取方法,有效地对高维步态轮廓特征进行降维,再利用BP神经网络进行特征分类识别。实验结果表明,算法达到了较高的识别率。
Gait recognition is an emerging field of biometric identification technology,it depends on the indi-vidual characteristics of people walk for identification.With non aggressive,hard to hide,on the system reso-lution,the advantages of low requirement for remote identification.Has been based on the visual analysis of human movement research hot spot.This paper presents based on principal component analysis feature ex-traction method.Effective dimensionality reduction of high-dimensional gait contour feature.Using a BP neu-ral network for feature classification.The experimental results show that,the algorithm achieves a higher recognition rate.
出处
《电子质量》
2014年第3期83-85,共3页
Electronics Quality
关键词
图像预处理
PCA
特征提取
BP神经网络
Image preprocessing
Principal component analysis
Feature extraction
BP neural network
