基于Gabor小波的人脸表情特征提取研究

为了使计算机能更好的识别人脸表情,对基于Gabor小波变换的人脸表情识别方法进行了研究。首先对包含表情区域的静态灰度图像进行预处理,包括对确定的人脸表情区域进行尺寸和灰度归一化,然后利用二维Gabor小波变换提取脸部表情特征,使用快速PCA方法对提取的Gabor小波特征初步降维。再在低维的空间中,利用Fisher准则提取那些有利于分类的特征,最后用SVM分类器进行分类。实验结果表明,上述提出的方法比传统的方法识别速度更快,能达到实时性的要求,并且具有很好的鲁棒性,识别率高。

Facial expression recognition based on Gabor wavelet transform

In order to make the computer have a better recognition to face expression,the method of facial expression recognition based on Gabor wavelets transform is discussed.Firstly,with pre-processing is executed to a given static grey image containing facial expression information.Pre-processing including the identification of pure face facial expression region,size and gray-scale normalized,the methods based on two-dimensional Gabor transform for feature extraction and fastPCA mentioned in this paper for diminishing Gabor feature are discussed.Secondly,in the low dimensional space,use the FLD to obtain the features useful to classification.Finally,SVM is applied to sort the facial expressions.Compared with the conventional methods,experimental results show that this method has fast identification speed and better higher recognition accuracy.