Seasons Of Code

Blind Source Separation    • Riddhish Bhalodia   

Blind Source Separation

Blind Source Separation

The project involves exploring various implementations of Independent Component Analysis on sound/images and demonstrating through an ipython notebook.

We explored two different approaches to the problem of separating sources of audio signals by minimizing the statistical dependence - fastICA and FOBI. A detailed explanation of the theory and the code can be found on the rendered notebook here. You can download and use the code and dataset from the repository. The notebook can also be found on the official iPython Wiki.