Abstract: This paper presents an algorithm of face recognition based on a LBP（Local Binary Pattern） operator and Isomap.It uses an [ε]-LBP operator to extract texture features of face images，and utilizes Isomap which is a non-linear data dimensionality reduction method to reduce the data dimensionality，the nature of available data structures is extracted.Let the nature data structures as inputting for the classifier of the face classification.The experimental results show that this algorithm is able to face a good image classification and recognition，especially in the case of small samples.

Key words: LBP operator, Isomap, texture feature, face recognition, data reduction

摘要： 提出了一种基于LBP算子和Isomap相结合的人脸图像识别算法。利用[ε]-LBP算子提取人脸图像纹理特征，然后用Isomap对高维的纹理特征进行数据降维，得到人脸数据的本质几何结构。最后将降维后的数据作为分类器的输入进行人脸分类识别。实验结果表明，该算法能够对人脸图像进行良好的分类识别，尤其是小样本的情况下。

关键词: LBP算子, 等距映射, 纹理特征, 人脸识别, 数据降维