摘要
提出了一种表情识别的新方法,首先通过基于小波变换的图像分解和K-L变换等处理,来抽取面部表情区域的有效鉴别特征,之后采用神经网络集成技术对六种典型表情进行识别。在CMU表情数据库上的实验表明,该方法达到了很高的识别率,而且对光照变化也有一定的不敏感性。
This paper introduces a novel method for human facial expression recognition. We first acquire the low-dimension features by using wavelet transformation and K-L transformation, then an neural network ensemble was construeted to recognize six basic facial expression. Experiments on CMU database show that it performs better than the single neural network, and does not sensitive to the lighting condition.
出处
《微电子学与计算机》
CSCD
北大核心
2006年第7期143-146,149,共5页
Microelectronics & Computer
关键词
人工心理
表情识别
小波变换
神经网络集成
Artificial psychology, Facial expression recognition, Wavelet transformation, Neural network ensemble
