CAJ | 학술논문

奇异特征区域在人脸直方图匹配过程具有更高的可分辨性。基于此，在人脸直方图特征匹配算法的基础上提出一种基于缩放因子的人脸直方图特征距离度量方法。该方法通过缩小直方图距离较近区域的距离、放大直方图距离较远区域的距离，并在匹配中突出高可分辨性区域，以获得更高的识别精度。实验表明该方法可以显著提高各种局部特征提取方法和直方图距离度量算法在人脸识别中的精度。
기이특정구역재인검직방도필배과정구유경고적가분변성。기우차，재인검직방도특정필배산법적기출상제출일충기우축방인자적인검직방도특정거리도량방법。해방법통과축소직방도거리교근구역적거리、방대직방도거리교원구역적거리，병재필배중돌출고가분변성구역，이획득경고적식별정도。실험표명해방법가이현저제고각충국부특정제취방법화직방도거리도량산법재인검식별중적정도。
Singular feature area in face histogram matching process has higher resolution .Based on this , the face of histogram fea-ture matching algorithm is proposed on the basis of a face histogram characteristic distance measurement method based on the scaling factor.The proposed method enhances the role of regions with more discriminative power in matching by shortening the histogram dis -tance between regions which are close to each other while enlarging the histogram distance between regions which are far to each other , which is enlightened by biologic vision and caricatures .Experiments show that the method can improve histogram matching precision of algorithms based on either different local feature extraction techniques or different histogram distance measures .