基于部分遮挡人脸识别算法的研究

Research on Face Recognition Algorithm Based on Partial Occlusion

研究人脸识别问题。针对当人脸采集的图像出现面部关键区域遮挡时,传统算法往往需要依靠面部主要关键特征进行识别,遮挡人脸的大部分特征消失,造成的误识别、漏识别问题。为解决上述问题,提出了基于遮挡人脸图片的识别方法。方法首先对遮挡人脸图像进行小波变换,然后建立特征粗糙集,根据特征加权融合算法将细节特征向量进行有效联系,进而根据联系性进行识别。实验结果表明,方法的能够对遮挡的人脸图像进行有效的识别,提高了身份识别的安全性和准确度。

Research face recognition problem. When face image region is occluded, the traditional algorithms often rely on facial key characteristics identification, causing face feature disappearance. This paper proposed an occluded face image based recognition method. Firstly, wavelet transform was carried out with the occluded face image. Then characteristics of rough sets were built up. According to the characteristics of weighted fusion algorithm, details feature vectors were contacted to recognize the face. Experimental results show that this method can recognize the occluded face image effectively, and improve the security and accuracy of identification.