jcwtlib-0.01.tar

  独立成分分析（Independent Component Analysis, ICA）是近年来发展起来的一种有效的盲分离技术，最早是由法国学者Herault和Jutten于1986年提出。ICA方法的提出最初是用来解决“鸡尾酒会”问题，其过程可以归纳为，在源信号与传输通道参数均未知的情况下，仅根据源信号的统计特性，出现测信号恢复出源信号。ICA分析的关键在于根据一定的优化准则建立描述输出信号独立程度的优化判据，即目标函数，并设计相应的优化算法，寻求最优的分离矩阵，使得输出信号中各分量尽可能相互独立。采用ICA方法对信号进行盲分离时，可避免过多地受大能量噪声成分的影响，将部分小能量信号源识别出来。随着ICA技术的不断发展，目前ICA技术已经广泛应用于语音增强、音频识别、通信、医学信号处理、金融数据挖掘、图像消噪、数字水印、遥感等领域，并取得了一定的研究成果。(Independent Component Analysis (ICA) Component ini order, developed in recent years is a kind of effective separation technology, originally from French scholars Herault and Jutten in 1986. The proposed method is initially ICA to solve the "cocktail" problem, the process can be divided into, in the source signal transmission channels and in the case of unknown parameters according to the source of the signals, and the statistical characteristics of recovering the source signal. According to the analysis of the ICA key lies in the establishment of optimal criteria described the output signal of the independent, objective function optimization criterion, and the corresponding optimization algorithm, the optimum )