摘要
提出了一种基于BP人工神经网络的人脸识别新算法。采用积分投影与几何特征提取相结合的方法进行人脸图像特征提取,构建特征向量,利用BP神经网络分类识别。仿真结果表明,该算法应用于ORL人脸库的分类识别,仅用13个特征即可达到平均识别率99%,识别能力显著增强,同时有效地降低了所需特征维数和计算复杂度。
New method,back propagation neural network based face recognition was presented in this paper.The proposed method extracts feature from face image with differential projection and geometrical features into eigenvector which was classified by back propagation neural network.The experimental results on ORL face database show that the proposed method can achieve an average recognition accuracy of 99% by using only 13 features.Moreover,the identified power was enhanced effectively,and the computing complexity and feature dimensions were reduced greatly.
出处
《液晶与显示》
CAS
CSCD
北大核心
2011年第6期836-840,共5页
Chinese Journal of Liquid Crystals and Displays
基金
国家自然科学基金资助项目(No.60772153)
关键词
人脸识别
BP神经网络
图像特征向量
face recognition
back propagation neural network
image eigenvector
