基于改进固定点ICA算法的图像盲分离

Blind Separation Based on Improved Fixed-point ICA Learning Algorithm

独立向量分析根据信源统计独立特性对观测信号进行分离运算,目前采用较多的是固定点独立分量分析(FastICA).考虑到图像信号分离中,图像信号复杂多样,信息量大的特点,采用改进固定点ICA算法对图像进行分离,克服了采用固定点ICA算法计算量大、收敛速度慢的缺点.文章采用随机提取的独立图像做实验,取得了稳定性较强的效果.

Signals were separated through Independent Component Analysis（ICA） based on independenees of the observed signal. Fixed-point independent Component Analysis algorithm is widely used nowadays. Considering complexity, diversity and much information of figure signal, an improved Fast ICA algorithm was used to separate the image signal, overcoming the large amount of calculation and the slow convergence of Fixed ICA algorithm. Some expefirnents were done with random images and achieved the stability of the results.