线性判别分析的红外光人脸表情识别

Facial expression recognition based on linear discriminant analysis in visible and near infrared light

人脸表情识别是人工智能领域一个重要的研究方向,当前方法在对人脸进行识别的过程中存在算法复杂、识别率低等不足。提出一种基于线性判别分析的可见光和近红外光人脸表情识别方法研究,首先对原始人脸表情样本进行下采样处理,将提取的测试样本表示为多个训练样本的线性组合;以矩阵指数为基础,对类间离散度和类内离散度进行重新定义,同时提取类内离散度矩阵的空间特征信息,实现对人脸表情的有效识别。实验证明,提出的方法能够有效提高运算速度、提高识别准确。

Facial expression recognition is an important research direction in the field of artificial intelligence. In the process of face recognition, the algorithm is complex and the recognition rate is low. A linear discriminant analysis method of visible light and near infrared face recognition was proposed. Based on the original face samples under sam- pling, test samples will be expressed as a linear combination of a number of training samples. In a matrix based index, redefine the between-class scatter and within-class scatter. At the same time, extrac spatial teature information from within class scatter matrix to achieve effective recognition of facial expression. Experimental results show that the pro- posed method can effectively improve the computing speed and recognition accuracy.