Therefore, the signal characterized

Therefore, the signal characterized with dictionary atoms can be extracted by step by step. In fact, the signal decomposition of MP is
the process of orthogonal projection. And that can ensure the decomposition’s uniqueness of chosen dictionary atoms. Fig. 4 shows the process of signal decomposition. From the process of decomposition based on MP, it is known that decomposing every times signal or residual signal must select the atom form over-completed dictionary, calculate and obtain the maximum value. In other words, every step of the calculation should be done based on Eq. (3) and changed to the multiple inner product calculation in a high-dimensional. Ref. [11], so this is the main reason that amount of calculation of signal decomposition is large. If take Eq. (3) as objective optimization function and use DEPSO algorithm, the amount of calculation will be dropped down obviously.

結果 (中文) 1:

因此，具有字典原子可以将信号提取由一步一步走。事实上，MP 的信号分解是正交投影的过程。并确保空房的字典原子分解的唯一性。图 4 显示了信号分解的过程。从基于 MP 分解的过程，众所周知的分解每次信号或残余信号必须选择原子形式超额完成的词典、 计算和获得的最大值。换句话说，每一步计算的应做基于情商 (3) 和改为在高维多的内积计算。文献 [11]，所以这是主要原因这一数额的计算信号的分解是大。如果以情商 (3) 为目标优化功能和使用 DEPSO 算法，计算量将明显下降。

結果 (中文) 3:

因此，可以提取的信号特征与字典原子可以通过一步一步来提取。事实上，国会议员的信号分解是正交投影法。这就保证了选择字典原子的唯一性。图4显示了信号分解过程。从分解的过程中，它是已知的每个时间的信号或残余信号的分解，必须选择的原子形式完成字典，计算和获得的最大值。换句话说，每一步的计算应根据公式（3）进行了多个内积计算高维。参考文献[ 11 ]，所以这是主要的原因，信号分解计算量大。如果将式（3）为优化目标函数，用DEPSO算法，计算量将明显下降。