Home My Page Projects MPTK: The Matching Pursuit ToolKit
Summary Activity Forums Tracker Lists Docs News SCM Files

Forum: help

Monitor Forum | Start New Thread Start New Thread
RE: Changing the cost function [ Reply ]
By: RĂ©mi Gribonval on 2014-09-30 16:07
[forum:148952]
Hi Branislaw,
If your goal is to use Gabor atoms choosen according to a new criterion, you may want to follow the developer manual to create a new type of "block" inheriting from Gabor blocks, with its own update_frame() method (this is where, if my memory is correct, correlations are computed) and create_atom() method to generate an atom with coefficients matching your new correlation.
I hope this helps.
Remi.

Changing the cost function [ Reply ]
By: Branislav Gerazov on 2014-09-24 16:27
[forum:148933]
Hi guys,

I'm using MPTK for some speech analysis and I was wandering - is it possible to change the cost function that MPTK minimizes? In other words I want to plug in my own function instead of the correlation used at the moment.
If so, can you point me to where in the code I can do this.

Thanks,
Branko

--
Branislav Gerazov
Faculty of Electrical Engineering and Information Technologies,
University of Ss. Cyril and Methodius - Skopje,
Macedonia