特征提取算法中利用样本间的协同表示关系构造邻接图只考虑所有训练样本的协同能力,而忽视了每一类训练样本的内在竞争能力。为此,本文提出一种基于竞争性协同表示的局部判别投影特征提取算法(competitive collaborative repesentation-based local discrininant projection for feature extraction,CCRLDP),该算法
In this note, using the discrete Calderon type reproducing formula, we prove the Plancherct-Poly a type inequality with the minimum regularity. As a consequence