基于特征融合的面部表情识别

Face expression recognition based on feature fusion

针对传统Gabor变换在提取表情特征时,冗余较大、特征维数较高的不足,结合ASM自动特征定位技术,提出了一种基于特征点Gabor特征和ASM形状特征相融合的面部表情识别方法,实验表明,两种特征的融合,可有效地利用特征点的局部纹理信息和脸部器官的整体形状信息,达到了更好的面部表情识别效果。

Aiming at the lack of the traditional Gabor transform in bigger redundancy and higher dimension, combining the ASM automatic feature orientation technology, the face expression recognition method based on the feature fusion between the Gabor feature and the ASM shape feature is proposed. The experiments indicate that the fusion of the two features can effectively uses the local texture information of feature dots and the holistic shape information of shape features, and obtains better recognition effect.