Ticket #31 (new task)
Julius - dealing with false-positives
Reported by: | kmaclean | Owned by: | kmaclean |
---|---|---|---|
Priority: | minor | Milestone: | Acoustic Model 1.0 |
Component: | Acoustic Model | Version: | 0.1-alpha |
Keywords: | Cc: |
Description
Need to understand how Julius returns recognition results so we can figure out how to flag 'out of grammar errors' - at first glance, it seems like this needs to be done in the application layer (rather than at the Julius layer).
May just need more speech audio data to train Julius with ...
Change History
Note: See
TracTickets for help on using
tickets.
May need to look at running a Statistical Acoustic Model in parallel with the Julian Grammar.