一种能实现免脱帽人脸识别系统算法

针对人脸识别系统在人脸被遮挡情况下识别率低的问题,为进一步提升人脸在遮挡情况下的识别率,文章提出一种通过图像多方向梯度值,使用融合、补偿等方式产生可以对原图像进行特征描述的特征图像,通过对特征图进行一系列处理后实现人脸识别的算法;算法首先计算图像四方位的梯度值;其次对4个梯度值进行融合运算,产生合融梯度、差融梯度;再次以合融梯度、差融梯度作为补偿变量在原图像上进行适当系数的补偿,形成人脸图像特征图;然后对特征图依次进行直方图统计、主成分分析后,使用SVM分类器进行分类识别;使用Matlab2016试验仿真平台在ORL、CMU_PIE等多个人脸数据库上进行测试,分别取得100%、92.21%的准确率,结果表明推荐算法在人脸被遮挡情况下的识别率具有很好的表现。

A face recognition system for identification without taking off a hat

Identification system for ship workers face recognition algorithm in the problem of low recognition rate of face obscured case, put forward a kind of gradient value by using image multiple directions, using the fusion algorithm and compensation methods such as image produced the original image can be described characteristics, based on the characteristics of the figure of a series of processing so as to realize the algorithm of face recognition.Firstly, the gradient value of image quadrangle is calculated.Secondly, four gradients are fused to produce convergence gradient and differential gradient.Thirdly, the fusion gradient and the difference gradient are used as compensation variables to compensate the appropriate coefficients on the original image to form the face image feature map.Then, histogram statistics and principal component analysis were performed on the feature map in turn, and SVM classifier was used for classification and recognition.Matlab2016 experimental simulation platform was used to test on ORL database. The results show that the algorithm presented in this paper has a good performance in face recognition when the face is obscured.