Is there an easy and efficient way to return a list of
recognized phrases together with some measure of
confidence in each one when running in "live" mode?
The reasons I am concerned that this could be
problematic are listed below:
1. In fbs.h it states:
float conf; /* Confidence measure (roughly
prob(correct)) for this word;
NOT FILLED IN BY THE
RECOGNIZER at the moment!! */
2. In sphinx2.html it states (where it describes -nbest):
No. of N-best hypotheses to be produced. Currently,
this flag is only useful in batch mode. But an application
can always directly invoke search_get_alt to obtain them.
Also, the current implementation is lacking in some
details (e.g., in returning detailed scores).
Thanks in advance.
John
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Is there an easy and efficient way to return a list of
recognized phrases together with some measure of
confidence in each one when running in "live" mode?
The reasons I am concerned that this could be
problematic are listed below:
1. In fbs.h it states:
float conf; /* Confidence measure (roughly
prob(correct)) for this word;
NOT FILLED IN BY THE
RECOGNIZER at the moment!! */
2. In sphinx2.html it states (where it describes -nbest):
No. of N-best hypotheses to be produced. Currently,
this flag is only useful in batch mode. But an application
can always directly invoke search_get_alt to obtain them.
Also, the current implementation is lacking in some
details (e.g., in returning detailed scores).
Thanks in advance.
John