摘要
为提高人脸识别的准确度,提出了一种基于低分辨率局部二值模式的人脸识别方法。该方法将原始人脸图像滤波下采样处理成低分辨率图像,将其划分成若干块矩形块图像,对每一块图像进行局部二值模式计算,统计出每一块LBP图谱的直方图,再连接在一起成为这幅图片的最终特征向量。经实验表明,该算法在ORL和YALE上均取得了更好的识别效果,且对光照、表情、姿势等的变化具备鲁棒性。
In order to improve the accuracy of human face recognition, a novel Low Resolution Local Binary Pattern （LRLBP） method is proposed. First, the low-resolution image is obtained from the original image through wave filter and down-sampling. Second, the obtained image is divided into nonoverlapping rectangular regions and the histogram of each region is computed independently. Finally, these histograms are concatenated as the feature vector of the human face. Experiment results on ORL and Yale face databases show that the proposed LRLBP method can get high human face recognition rate and that the method is robust to illumination, face expression and position variations.
出处
《吉林大学学报（工学版）》
EI
CAS
CSCD
北大核心
2013年第2期435-438,共4页
Journal of Jilin University:Engineering and Technology Edition
基金
国家自然科学基金项目(61101155)
吉林省科技发展计划项目(20101504)
关键词
计算机应用
局部二值模式
低分辨率
特征提取
人脸识别
computer application
local binary pattern
low resolution
feature extraction
face recognition
