CAJ | 학술논문

LBP（局部二值模式）作为一种有效的纹理描述算子，度量和提取图像局部的纹理信息，对光照具有不变性，在单幅人脸图像识别具有很好的应用。在研究此理论的基础上提出了一种基于局部二值模式（LBP）与二维离散余弦变换（2DDCT）相结合的人脸识别方法。将单幅的人脸图像规则的分成多个子块，对每个子块进行LBP变换，把每个LBP变换的后的子块分别用2DDCT变换到频率域，大部分信息保存在低频部分，选取低频作为人脸的频率域特征，有效的降低了维数，使计算量降低。实验结果表面，在ORL人脸数据库上具有较高的识别率。
LBP（국부이치모식）작위일충유효적문리묘술산자，도량화제취도상국부적문리신식，대광조구유불변성，재단폭인검도상식별구유흔호적응용。재연구차이론적기출상제출료일충기우국부이치모식（LBP）여이유리산여현변환（2DDCT）상결합적인검식별방법。장단폭적인검도상규칙적분성다개자괴，대매개자괴진행LBP변환，파매개LBP변환적후적자괴분별용2DDCT변환도빈솔역，대부분신식보존재저빈부분，선취저빈작위인검적빈솔역특정，유효적강저료유수，사계산량강저。실험결과표면，재ORL인검수거고상구유교고적식별솔。
LBP (Local binary pattern) as an effective measure and describe a partial extract image texture information operator, which is invariant to illumination and that has good applicaton in single human face recognition. In this paper, based on the study of this one theory, we propose a new face recognition method which is on the basis of integrating the local binary pattern(LBP) with two-dimensional discrete cosine transform (2DDCT). The face was divided regularly into small regions, for each sub-block transformation to LBP and then transformation to the frenquency domain by 2DDCT. Most of the information stored in the low frequency section and s-Elect low frequency as the frequency domain characteristics of the human face, which is reducing the number of dimensions ecffectively, so that reduces the amount of calculation. The experimental results show that this method has a high recognition rate on ORL face databases.