CAJ | 학술논문

针对基于局部模式的人脸识别方法特征维数高、计算复杂度高、识别时间长的问题，提出一种结合主成分分析和局部导数模式的人脸识别方法，并针对如何解决光照、人脸表情等方面的问题提出了改进的编码方法。该方法首先将人脸图像分成很多小的区域，然后在每一个小区域中用改进的编码方法进行编码，并建立该区域的局部导数直方图，然后采用主成分分析法对所有直方图向量进行降维得到特征向量，最后利用最近邻分类器计算相似度。实验表明，这里提出的结合主成分分析和局部导数模式方法无论在识别率还是在运算速度上都优于传统的识别算法。
침대기우국부모식적인검식별방법특정유수고、계산복잡도고、식별시간장적문제，제출일충결합주성분분석화국부도수모식적인검식별방법，병침대여하해결광조、인검표정등방면적문제제출료개진적편마방법。해방법수선장인검도상분성흔다소적구역，연후재매일개소구역중용개진적편마방법진행편마，병건립해구역적국부도수직방도，연후채용주성분분석법대소유직방도향량진행강유득도특정향량，최후이용최근린분류기계산상사도。실험표명，저리제출적결합주성분분석화국부도수모식방법무론재식별솔환시재운산속도상도우우전통적식별산법。
In order to solve the problems of high feature dimension,high computational complexity and long recognition time caused by the face recognition method based on local pattern,a face recognition method combining principal component analysis and local derivative model is presented in this paper. The improved coding method is introduced to solve the problems of illumination,facial expression and so on. The face image is divided into many small regions,and then each small area is coded by the improved coding method and a LDP histogram of the region is created. The dimension of histogram vectors are reduced by using principal component analysis. Finally,the nearest neighbor classifier is used to calculate the similarity. Experimental re-sults show that the presented method is superior to the traditional method in the recognition rate and the speed.