liao0009
Newbie level 5
blind source separation tutorial
Anyone work in convolutive mixture Blind source separation problem? I am quite interested in this topic, and trying to implement a simple BSS in Matlab under convolutive mixture. But i have a confusion abt the "averaging operation" proposed in the learning algorithms, for example, in the paper attached below, authors propose the learning algorithm in frequency domain:
W(r+1,m) = W(r,m) + eta * [ I - <phi(Y(r,m))Y'(r,m))>] * W(r,m), where eta is the step size, phi is a non-linear function which is defined in the paper, and < . > is an averaging operation, which appears many times in most of the learning algorithm of convolutive BSS.
does anyone know how this averaging operation works?
A short tutorial about BSS can be found online at:
**broken link removed**
Anyone work in convolutive mixture Blind source separation problem? I am quite interested in this topic, and trying to implement a simple BSS in Matlab under convolutive mixture. But i have a confusion abt the "averaging operation" proposed in the learning algorithms, for example, in the paper attached below, authors propose the learning algorithm in frequency domain:
W(r+1,m) = W(r,m) + eta * [ I - <phi(Y(r,m))Y'(r,m))>] * W(r,m), where eta is the step size, phi is a non-linear function which is defined in the paper, and < . > is an averaging operation, which appears many times in most of the learning algorithm of convolutive BSS.
does anyone know how this averaging operation works?
A short tutorial about BSS can be found online at:
**broken link removed**