Abstract: The accuracy of face recognition can not satisfy the researchers and engineers for pose invariance，lighting condition and expression.This paper proposes a novel method to present the facial data and improve the accuracy，which takes into both feature information and structure information.Experiments show that the proposed method is effective to expression recognition with AR and ORL datasets，and is robust to lighting and expression variable.

Key words: face recognition, image partition, Linear Discriminant Analysis（LDA）

摘要： 人脸识别由于受到光照、表情、姿势等因素的影响一直没有能够取得非常满意的识别率。提出了基于图像分割的人脸识别方法，该方法同时考虑了人脸图像中特征信息和结构信息，实验采用AR和ORL人脸数据库，结果表明该方法能有效提高人脸识别的准确率，同时对表情变化、灯光具有一定的鲁棒性。

关键词: 人脸识别, 图像分割, 线性判别分析