在目前的人脸表情识别系统中，人脸表情的机器识别和人类感知之间存在着本质的差异，造成人脸表情识别率不高。为了减小人脸图像底层视觉特征与高层语义之间的语义鸿沟，提出一种基于层次分析法（AHP）语义知识的人脸表情识别新方法。该方法首先采用层次分析法对训练集中人脸图像进行高层语义描述，建立语义特征向量，在底层视觉特征提取阶段，提出一种二阶PCA（principal component analysis）方法来提取人脸图像的纹理特征；在识别阶段，仅利用输入人脸图像的底层视觉特征，采用K-NN（k-nearest neighbor）算法并结合学习阶段建立的语义特征向量，进行人脸表情分类识别。提出的人脸表情识别方法结合底层视觉特征和高层语义知识，减小了人脸图像底层视觉特征与高层语义之间的语义鸿沟。在JAFFE人脸表情数据库中进行实验，获得了93.92%的平均识别率。理论分析和实验结果表明，与其他的人脸表情识别方法相比，该方法具有更好的识别效果。

At present，there are intrinsic differences between machine recognition of facial expression and human perception in the facial expression recognition system, which affect the precision of facial expression recognition. In order to reduce the semantic gap between the low-level visual features of face images and high-level semantic, a novel facial expression recognition method based on semantic knowledge of analytical hierarchy process(AHP) is presented. The analytical hierarchy process method is adopted to describe the high-level semantic of face images of the training set, which further used to establish semantic features.In the stage of low-level visual features extraction, the 2nd-order principal component analysis method is proposed to extract the texture features of face images. In the recognition stage, only low-level visual features of the input face image is used, and k-nearest neighbor method combined with semantic features in the study stage is used to classify the facial expressions. The proposed method combines the low-level visual features with high-level semantic features, reducing the semantic gap between them.The experiments are conducted on Japanese Female Facial Expression (JAFFE) database and the overall recognition rate of 9392% is achieved. Theoretical analysis and experimental results both show that the proposed method has higher recognition rate.