摘要
传统算法进行模糊人脸识别的过程中,一旦人脸表情发生变化,人脸特征也将发生改变,导致人脸识别的准确性降低。为此,提出一种基于改进的格拉斯曼流形的模糊人脸识别方法。在格拉斯曼流形上构建全部模糊人脸样本图像的近邻图来估计人脸特征分布的几何结构,然后将其作为正则化项整合到模糊人脸识别的目标函数中,从而获得更精确的人脸特征投影矩阵。仿真实验结果表明,利用改进算法进行模糊人脸识别,能够提高识别的准确率和效率,效果令人满意。
The accurate recognition for fuzzy face has important application value and practical significance. In the process of fuzzy face recognition, since the traditional algorithm has the shortcoming of low accuracy of face recognition, a fuzzy face image recognition algorithm based on the improved Grassmannian. is proposed. The neighbor graphs of all fuzzy face sample images are constructed on Grassmannian to estimate the geometric structure of the facial features, and then it is integrated into the objective function as the regularization term to obtain the accurate projection matrix of the facial features. The simulation experiment results show that the improved algorithm for fuzzy face recognition can improve the recognition accuracy and efficiency. The results are satisfactory.
出处
《现代电子技术》
北大核心
2015年第22期34-36,40,共4页
Modern Electronics Technique
基金
佛山市产学研专项资金项目:基于人脸识别的企业考勤系统(2012HC100303)
关键词
改进的格拉斯曼流形
模糊人脸识别
人脸特征分布
人脸识别方法
improved Glassmannian
fuzzy face identification
face feature distribution
face identification method