中文摘要：

目前多数人脸表情识别的研究仅限于6种基本表情，未考虑到人脸表情变化是细微的。因此提出了基于混合特征和分类树的细微表情识别方法。对眼睛区域采用Gabor小波变换提取纹理变化特征，对鼻子区域采用2D—DCT提取纹理变化特征，而对嘴巴区域采用改进的AAM提取形状变化特征。分类识别时，将易混淆表情先归为一类进行表情的粗分类，然后对类内的表情选择相应表情贡献较大的特征子区域中的特征，进行表情细分类。在每级分类识别过程中，对每个区域采用离散HMM得出表情概率，最后采用在训练阶段得到的贡献权值进行加权融合得到分类结果。实验结果表明，该方法能够得到较好的识别效果，且处理速度快，适合于实时图像序列的细微表情识别。

英文摘要：

Most of facial expression recognition researches are focused on six basic facial expressions currently, and subtle facial expressions recognition has not been considered. A method of subtle expression recognition based on hybrid features and classifier tree is presented in this paper. Texture feature for the eye area is extracted by using Gabor wavelet transformation, texture feature for the nose area is extracted by using 2D-DCT, and shape variety feature for the mouth area is extracted by using AAM. At the recognition stage, the expressions are classified roughly by subsuming confusable expressions in the same kind of expression firstly. Then the expression features that contribute much to this kind expression are selected for fine classification. In the process of each grade classification, discrete HMM is adopted for expression recognition in each expression area respectively. The classification results are fused by means of integrating the probability of each expression in each area with its weight obtained during the training phase. Experiments show that this method can get high recognition rate. And it also has high speed and is suitable for real time subtle expression recognition.