Next: 13.14.3 Tracing
Up: 13.14 HResults
Previous: 13.14.1 Function
HRESULTS is invoked by typing the command line
HResults [options] hmmList recFiles ...
This causes HRESULTS to be applied to each recFile in turn.
The hmmList should contain a list of all model names for
which result information is required. Note, however, that
since the context
dependent parts of a label can be stripped, this list is
not necessarily the same as the one used to perform the
actual recognition.
For each recFile, a transcription file with the same name
but the extension .lab (or some user specified extension - see
the -X option) is read in
and matched with it. The recfiles may be master label files (MLFs).
The available options are
- -a s
- change the label SENT in the output to
s.
- -b s
- change the label WORD in the output to
s.
- -c
- when comparing labels convert to upper case. Note that
case is still significant for equivalences (see -e below).
- -d N
- search the first N alternatives for each test label
file to find the most accurate match with the reference labels.
Output results will be based on the most accurate match to allow
NBest error rates to be found.
- -e s t
- the label t is made equivalent to the
label s. More precisely, t is assigned to an equivalence
class of which s is the identifying member.
- -f
- Normally, HRESULTS accumulates statistics for all
input files and just outputs a summary on completion.
This option forces match statistics to be output for each
input test file.
- -g fmt
- This sets the test label format to fmt.
If this is not set, the recFiles should be in the
same format as the reference files.
- -h
- Output the results in the same format as US NIST scoring
software.
- -k s
- Collect and output results on a speaker by speaker basis
(as well as globally). s defines a pattern which is used
to extract the speaker identifier from the test label file name.
In addition to the pattern matching metacharacters
* and ?
(which match zero or more characters and a single character
respectively), the character % matches any character whilst
including it as part of the speaker identifier.
- -m N
- Terminate after collecting statistics from the first
N files.
- -n
- Set US NIST scoring software compatibility.
- -p
- This option causes a phoneme confusion matrix to
be output.
- -s
- This option causes all phoneme labels with the form
A-B+C to be converted to B. It is useful for
analysing the results of phone recognisers using context dependent
models.
- -t
- This option causes a time-aligned transcription of
each test file to be output provided that it differs from
the reference transcription file.
- -u f
- Changes the time unit for calculating false alarm
rates (for word spotting scoring) to f hours (default is 1.0).
- -w
- Perform word spotting analysis rather than string
accuracy calculation.
- -z s
- This redefines the null class name to s.
The default null class name is ???, which may be difficult
to manage in shell script programming.
- -G fmt
- Set the label file format to fmt.
- -I mlf
- This loads the master label file mlf. This option
may be repeated to load several MLFs.
- -L dir
- Search directory dir for label files (default
is to search current directory).
- -X ext
- Set label file extension to ext
(default is lab).
HRESULTS also supports the standard options -A,
-C, -D, -S, -T, and -V as described
in section 4.4.
Next: 13.14.3 Tracing
Up: 13.14 HResults
Previous: 13.14.1 Function
ECRL HTK_V2.1: email [email protected]