基于二维判别局部排列的特征提取算法

Feature Extraction Algorithm Based on Two-dimensional Discriminant Locality Alignment

提出一种基于二维判别局部排列的特征提取算法,用于人脸识别等模式分类问题。该算法集成局部判别排列和二维特征提取算法的思想,在部分优化阶段,对每一个训练样例,通过寻找最近邻的方式构建局部面片,设计目标函数保持局部判别信息。在整体排列阶段,利用排列技术,整合各个局部面片,得到一个全局的坐标系。通过求解特征值问题可得到低维投影矩阵。在ORL标准人脸数据库上的实验结果表明,该算法具有较好的优越性和鲁棒性。

Anew subspace learning algorithm,which is called Two-dimensional Discriminant Locality Alignment（2DDLA） algorithm,is proposed for pattern classification,such as face recognition.The proposed algorithm integrates the idea of discriminant locality alignment and two-dimensional feature extraction algorithm.2DDLA operates in the following two stages： First,in part optimization stage,for each sample,it constructs local patch by seeking for the nearest neighbors.An object function is designed to preserve local discriminant information.Second,in whole alignment,the alignment trick is used to align all part optimizations to the whole optimization.The projection matrix can be obtained by solving a standard eigen-decomposition problem.Experimental results on ORL face database show that the algorithm has better superiority and robustness.