Library Name: lib_pr.a
Introduction:
Pattern Recognition - the act of taking in raw data and taking an action
based on the category of the pattern - plays a central role in speech
recognition. The Pattern Recognition library consists of classes which
will handle parametric, non-parametric, stochastic and non-metric
techniques for both supervised and un-supervised learning. The IFC's contain the following Pattern Recognition classes.
Example:
01 // file: $isip/doc/examples/class/pr/pr_example_00/example.cc
02 // version: $Id: index.html 10298 2005-12-13 20:46:47Z may $
03 //
04
05 // isip include files
06 //
07 #include <HiddenMarkovModel.h>
08
09 int main(int argc, const char** argv) {
10
11 // declare a HiddenMarkovModel object
12 //
13 HiddenMarkovModel hmm;
14
15 // read the hmm parameters from the parameter file
16 //
17 Filename param_file(L"$ISIP_DEVEL/class/pr/HiddenMarkovModel/diagnose_params.sof");
18 Sof param_sof;
19
20 param_sof.open(param_file);
21 hmm.read(param_sof, param_sof.first(hmm.name()));
22 hmm.setVerbosity(Integral::ALL);
23 param_sof.close();
24
25 // set up the input speech file list
26 //
27 Sdb raw_list_sdb;
28 boolean is_list = true;
29 raw_list_sdb.append(L"$ISIP_DEVEL/class/pr/HiddenMarkovModel/diagnose_raw_list.sof", is_list);
30
31 // set up the front end parameters file
32 //
33 hmm.setParamFile(param_file);
34
35 // run the HiddenMarkovModel recognizer on the input speech file list
36 //
37 hmm.run(raw_list_sdb);
38 }
Explanation:
We first read a HiddenMarkovModel object from the parameter file (line
21) and set up the verbosity level (line 22). Then we specify the
input file list (line 29). Finally we run the recognizer (line 37).
Note that there are many changes expected in this class.
The pattern recognition classes will include:
The software corresponding to the examples demonstrated in this document
can be found in our documentation directory
under class/pr/.