摘要
由于实际所观测信号常常掺杂有噪声,直接对其进行独立分量分析(Independent Component Analysis,ICA)将会影响ICA的分离性能.本文将Wiener滤波与ICA进行结合,先对有噪信号进行Wiener滤波,再进行ICA分离.应用上述算法,对有噪图像信号进行仿真实验研究,通过比较降噪前后的峰值信噪比(Peak Signal to Noise Ratio,PSNR),说明该算法的可行性和有效性.
Due to the fact of measurement data corrupted with additive noise in many practical applications, the separation performance of ICA （ Independent Component Analysis, ICA） will be influenced without de-noising. The aim of this paper is to demonstrate that wiener filter is extremely attractive for efficient source separation of noisy ICA mixtures. The simulated experiments of woisy images using ICA algorithms aften wiener de-noising are implemented and the PSNR（ Peak Signal to Noise Ratio）is used to provide quantitative evaluations of the feasibility and efficiency of the method.
出处
《南昌大学学报（工科版）》
CAS
2007年第1期87-90,共4页
Journal of Nanchang University(Engineering & Technology)
基金
国家自然科学基金资助项目(50475117)
关键词
WIENER滤波
独立分量分析
盲源分离
降噪
wiener filter
independent component analysis
blind source separation
de-noising
