基于ICA的混合图像盲分离算法性能初探

Study on the Performance of ICA-Algorithm Performed in Mixed Images Blind Separation

为了找出适用于混合图像分离的优秀算法，针对不同的ICA算法，使用基于Matlab的经典ICALAB软件包分别在相同混合矩阵．不同混合矩阵条件下对比了各算法在混合图像分离中的性能指标和迭代时间，对比了各算法在混合图像分离中的性能。最终，通过仿真实验对比得出了SANG和FPICA算法是适用于混合图像分离的优秀算法。

In order to find out excellent algorithms performed in mixed images separation, aim at different ICA algorithm , the performance index and elapse time of very algorithm are investigated under same mixed matrix and different mixed matrices using ICALAB based on Matlab. Through simtdations performance of very algorithm are compared. In conclusion, SANG and FPICA perform well in mixed image separation.