摘要
:从协同模式识别的动力学方程出发 ,研究了协同识别算法中的序参量和注意参数 ,提出了变注意参数方法 ,推导了相应的变化方程 ,以加快人脸识别过程 ,减少迭代次数 .对加噪人脸图像的识别实验表明 ,协同算法对噪声具有较好的鲁棒性 ,同时识别迭代次数减小至 1 /8左右 。
Face recognition is one of the important applications of pattern recognition. Synergetic algorithm has the advantage of recognizing objects by whole image information. Order parameters and attention parameters in synergetic algorithm were discussed through analyzing dynamic equation. Then, a new method of changing attention parameters and its equations were presented, which can lessen the recognition iteration process. The experiment using noised face images proves that the synergetic algorithm is robust and the new method can lessen iteration by about 8 times.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2000年第6期821-823,共3页
Journal of Shanghai Jiaotong University
关键词
人脸识别
模式识别
序参量
动力学方程
协同算法
synergetic
face recognition
dynamic system
order parameters
attention parameters
