摘要
介绍基于神经网络的盲源分离算法的原理和结构,针对基于最小互信息后馈神经网络,建立了新的盲源分离算法,比起简单的独立分量分析,该算法可以更有效地对混合图像进行分离,计算机仿真实验验证了该算法的优良性能。
Introduces the principles and structure of blind source separation algorithm based on neural network, against each other based on minimum information fed neural network, the establishment of a new algorithm for blind source separation, rather than simple independent component analysis, the algorithm can be more effective separation of mixed images, computer simulation proves that the algorithm＇s performance is excellent.
出处
《现代计算机》
2008年第10期99-101,共3页
Modern Computer
关键词
反馈神经网络
最小互信息
分离性能指数
信噪比
Feedback Neural Network
Minimum Mutual Information
Separation Performance Index
Signal-Noise Ratio
