本代码实现基于成对约束的半监督图嵌入算法-Following the intuition that the image variation of faces can be effectively modeled by low dimensional linear spaces, we propose a novel linear subspace learning method for face analysis in the framework of graph embeddi

eigen structure analisis

基于“脉动”结构 求 任意偶数阶实对称阵的特征值，用于MIMO通信反馈信道质量，已经在实际产品实现。( 注：脉动结构只需O(n*log(n))量级运算，高效快速)-"Systolic Arrays" structure based Eigen Value Decomposition