一种基于局部奇异值分解和最近邻决策规则的人脸识别方法

An Approach for Face Recognition Based on Local Singular Value Decomposition and Nearest Neighbor Decision Rule

提出了一种基于局部奇异值分解和最近邻决策规则的人脸图像识别方法。其主要内容包括以下方面:由于奇异值向量具有稳定性、转置不变性等特点,对归一化的人脸图像,采用局部奇异值分解抽取人脸图像特征作为识别特征;针对人脸识别问题,采用最近邻决策规则取代隶属度函数来进行分类识别。实验结果显示,所提出的方法减少了数据计算量,运行速度快,并提高了识别率。同时,人脸识别结果也证明了该方法的有效性。

A novel human face recognition method based on local singular value decomposition and nearest neighbor decision rule has been proposed, its essential contents can be listed as follows. With the stability and the invariability of rotation, local singular value decomposition is used to extract the features of the normalized face image. For face recognition problem, instead of using the membership functions method, nearest neighbor decision rule is used to classification and recognition. The results of experiments show that the proposed method reduces the computing time, and runs faster. Besides, the recognition rate is increased.