If the assumed pdf is different from the true pdf considerably, the sources will not be separated correctly.

• 盲源分离算法通常需预先假设源信号的概率密度函数,并由此获得关键的激活函数,进而从混合信号中分离出源信号。