Skip to content. Skip to navigation

The AMI Meeting Corpus is a multi-modal data set consisting of 100 hours of meeting recordings

Sections
Personal tools
You are here: Home Bug in word segmentation (time alignment) Re: Bug in word segmentation (time alignment)

Forum Post : Re: Bug in word segmentation (time alignment)

Up one level
Posted by jonathan at 2007-02-20 17:33

Thanks for pointing this out Igor. The main problem was in my algorithm for
taking forced alignment timings and putting them back into NXT format. I also
needed to re-run the forced-alignment process for all non-scenario meetings to
get improved results.


This problem and several others have been fixed by my new re-run of the forced-
alignment process for all non-scenario meetings. The fixed files are in CVS and will
appear in all future releases (from 1.3 onwards).

Jonathan Kilgour



Powered by Plone