基于核方法的相关自适应子空间分割  

Adaptive Subspace Segmentation Based on Kernel Method

作　　者：洪岚[1] 魏莱[1] 

机构地区：[1]上海海事大学信息工程学院,上海201306

出　　处：《现代计算机》2017年第16期35-38,42,共5页Modern Computer

摘　　要：人脸识别是模式识别研究的热门课题,有很重要的应用价值。针对在高维非线性人脸数据集的识别困难问题,提出一种基于核方法的相关自适应子空间分割方法(KCASS)。将相关自适应子空间分割(CASS)与核方法相结合,构成Kernel CASS方法。实验证明,该方法在效率上和识别效果上较之前的分类方法有进一步改善。Face recognition is a hot topic in pattern recognition, and it has important application value. In order to solve the problem of high dimensional nonlinear face recognition, proposes an adaptive subspace segmentation method based on kernel method （KCASS）. CASS and kernel method are combined to form the Kernel CASS method. The experimental results show that the proposed method is more efficient and effective than the previous classification methods.

关 键 词：人脸识别 核方法 子空间分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]