摘要
特征脸方法是人脸识别领域中的一种重要方法。本文在特征脸的基础上提出了一种新的方法,在特征脸空间,用向量之间的夹角来衡量图像之间的相似度,用最近邻法对图像进行分类和识别。我们用ORL提供的标准人脸库进行了测试,并与传统的特征脸方法进行了比较。结果表明,新方法的识别正确率明显高于传统特征脸法。
Eigenface is an important approach in the field of face recognition. In this paper, a new approach is presented on the basis of eigenface. It uses the angle between vectors of eigenface space as the similarity between face images, then recognizes and classifies face image by the nearest neighbor algorithm. In experiments on ORL face database, the new approach is compared to the traditional eigenface. The result shows that the new approach gets higher recognition rates than traditional one significantly.
出处
《微计算机信息》
北大核心
2006年第10S期297-298,285,共3页
Control & Automation
基金
国家发展计算委员会资助项目(19981303)
关键词
人脸识别
特征脸
夹角最小法
最近邻法
recognition,eigenface,miaidmal angle,the nearest neighbor algorithm
