CAJ | 학술논문

为了提高人脸识别的准确率，提出一种基于Gabor特征和鲁棒稀疏表征相融合的人脸识别算法（Gabor-RSC）。首先采用Gabor滤波器提取人脸图像的多尺度和多方向特征，并采用主成分分析降低特征维数，然后采用鲁棒稀疏编码算法对人脸进行识别，最后采用Yale和ORL人脸库进行仿真测试。结果表明，Gabor-RSC算法提高了人脸的识别正确率，鲁棒性更高。
위료제고인검식별적준학솔，제출일충기우Gabor특정화로봉희소표정상융합적인검식별산법（Gabor-RSC）。수선채용Gabor려파기제취인검도상적다척도화다방향특정，병채용주성분분석강저특정유수，연후채용로봉희소편마산법대인검진행식별，최후채용Yale화ORL인검고진행방진측시。결과표명，Gabor-RSC산법제고료인검적식별정학솔，로봉성경고。
In order to improve face recognition accuracy,the paper proposes a face recognition algorithm,it is based on the combination of Gabor feature and robust sparse representation (Gabor-RSC).First,it uses Gabor filter to extract multi-scale and multiple orientation features of face images,and employs principal component analysis to reduce the features dimensionality;then it uses robust sparse coding algorithm to recognise the face;finally,it conducts simulation tests on Yale and ORL face databases.Results show that the proposed Garbor-RSC algorithm improves face recognition accuracy,and has higher robustness.