摘要
提出一种基于图像关键块空间分布与Gabor滤波的人脸表情识别算法。该算法在传统的基于Gabor滤波的表情识别的基础上,增加表情图像的关键块空间分布信息,提高表情识别的准确率。首先,使用5个尺度8个方向的Gabor滤波器组对表情图像进行滤波,提取表情图像的Gabor特征;然后,使用人脸表情训练样本通过向量量化方法训练指定长度的码书,利用码书将训练样本图像编码成索引矩阵,获取表情图像的索引分布;最后,将图像编码获得的索引矩阵与Gabor特征共同作为表情图像的特征,用于表情识别。实验结果表明:本算法的识别效果比单独使用Gabor特征的表情识别要好。
A new facial expression recognition algorithm was proposed based on both image’s keyblock spatial distribution and Gabor filters.The keyblock distribution information was added to the conventional Gabor filter based expression recognition algorithm,which improved the accuracy of expression recognition.Codebook with specific length was trained by employing the vector quantization technique,and then each image in the train set was encoded into index matrix in order to extract the index distribution of the image.Meanwhile,Gabor filter bank with 5 scales and 8 orientations was used to extract Gabor features of expression images.Finally,Gabor features and keyblock distribution information were combined to be used for expression recognition.Experimental results show that the proposed algorithm achieves better results than the algorithm that only based on Gabor features.
出处
《中南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2013年第S2期239-243,共5页
Journal of Central South University:Science and Technology
基金
国家自然科学基金资助项目(61175096)
关键词
表情识别
关键块空间分布
GABOR滤波
码书
facial expression recognition
keyblock distribution
Gabor filter
codebook