摘要
针对人脸识别算法对光照变化敏感的问题,提出一种基于光照鲁棒稀疏表示的人脸识别方法。该方法对图像作小波变换,得到光照归一化图像,通过对光照归一化后人脸图像作稀疏变换,稀疏表示分类得出测试识别结果。本文方法在Yale B人脸库上仿真实验,识别率较高,对光照、表情、遮挡具有一定的鲁棒性。
To solve the problem that face recognition algorithm is sensitive to illumination change,a face recognition method based on illumination robust sparse representation is proposed.In this method,wavelet transform is applied to the image to get the illumination normalized image,and sparse transform is applied to the face image after the illumination normalized,and sparse representation classification is used to get the test recognition result.The method in this paper is simulated on Yale B face database,which has high recognition rate and certain robustness to illumination,expression and occlusion.
作者
朱强军
Zhu Qiangjun(Wanjiang College of Anhui Normal University,Wuhu Anhui,241000)
出处
《电子测试》
2020年第2期47-49,共3页
Electronic Test
基金
安徽省自然科学研究重点项目“基于人脸识别的智能门禁系统研究(KJ2018A0655)”
校级科研项目“基于人脸识别的门禁系统研究(WJKY-201723)”
关键词
人脸识别
稀疏表示
鲁棒性
face recognition
sparse representation
robustness
