中文说明：最近的K-SVD算法的学习过程中，采用稀疏编码部分OMP跟踪算法，尤其是对于本摘要。解决问题：其中D为过完备字典中，已经给出，Y是原始信号中，X的未知。其基本思想的OMP算法是：贪婪迭代方法来选择D列后面，以便与相关的当前冗余最大程度所选列向量的每个迭代中，减去从原始信号向量的相关部分，并多次迭代过程中，只有数迭代实现稀疏的K，停止迭代。
English Description：
In the recent learning process of K-SVD algorithm, sparse coding partial OMP tracking algorithm is used, especially for this summary. Solution: where D is given in the over complete dictionary, y is the unknown value of X in the original signal. The basic idea of OMP algorithm is: greedy iterative method to select the D column after, so as to maximize the redundancy with the relevant current. In each iteration of the selected column vector, subtract the relevant part from the original signal vector, and in multiple iterations, only a few iterations to achieve sparse K, stop the iteration.