rover.1 4.28 KB
.TH rover 1 "" "" "" ""
WORK IN PROGRESS 

NAME
rover - Recognition Output Voting Error Reduction
.PP
.PP

NOTE: This manual page was created automatically from
HTMl pages in the sclite/doc directory.  This manual page does not
include output file examples.  The author suggests using a HTML browser
for reading the sclite documentation.
.PP
SYNOPSIS
rover [ -sT -a alpha -c Nconf -f level -l width ] ( -h hypfile ctm )+ -o outfile -m meth
.PP
\*LInput Options \*O
.RS
-h hypfile ctm
.RS
Define the hypothesis file and it's format. This option
must be used more than once.  Currently, only the ctm format
is recgnized.
.RE
.RE
.PP
\*LOutput Options \*O
.RS
-o outfile

.RS
  Define the output file.  (Will be same format as hyps)
.RE
-f level

.RS
    Defines feedback mode, default is 1
.RE
-l width   

.RS
 Defines the line width.
.RE
.RE
.PP
\*LVoting Options \*O
.RS
-m meth 

.RS
 Set the voting method 'meth' to one of the following:

.br


.br

oracle   -> output the fully alternated transcript
.br

meth1    -> alpha = -a , conf = -c, choose highest avg
.br

maxconf  -> \*LVoting by Using Maximum Confidence Score\*O
.br

avgconf  -> \*LVoting by Average Conf. Score\*O
.br

maxconfa -> 
.RS
Same as maxconf, but set the confidence score for
the NULL transition arcs to be the number of NULL transition
arcs in the correspondence set divided by the number of input
systems.
.RE
putat    -> Output the putative hit format
.RE
-a alpha   
.RS
 Set Alpha to 'alpha'.  Alpha is the tradeoff between 
using word occurrence counts and confidence scores.
.RE
-c Nconf 

.RS
 Set confidence score associated with NULL transition
arcs to 'Nconf'.  Default: 0.0
.RE
.RE
.PP
\*LAlignment Options \*O
.RS
-s        

.RS
  Do Case-sensitive alignments.
.RE
-T  
.RS
  Use time information, (if available), to calculated word-to-
word distances.
.RE
.RE
.PP
DESCRIPTION
.PP
Rover is a tool combine hypothesized word outputs of multiple
recognition systems and select the best scoring word sequence. 
Rover is part of the \*LNIST SCTK\*O Scoring Tookit. A
number of different output formats can be generated and different
scoring functions can be specified.  A more complete description of
the rover system can be found in the paper \*L
A post-processing system to yield reduced word error rates: Recognizer
Output Voting Error Reduction (ROVER). 
.PP
 The ROVER system is implemented in two modules. First, the system
outputs from two or more ASR systems are combined into a single word
transition network. The network is created using a modification of the
dynamic programming alignment protocol traditionally used by NIST to
evaluate ASR technology. Once the network is generated, the second
module evaluates each branching point using a voting scheme, which
selects the best scoring word (with the highest number of votes) for
the new transcription.  The following figure depicts the the overall
system architecture.
.PP
The heart of the Rover program is the ability to combine system
outputs of mulitple recognition systems using an iterative Dynamic
Programming alignment protocol into a single, composite Word
Transition Network (WTN).  The protocol is fully described in the
\*L
Section 2.1. RECOGNITION OUTPUT ALIGNMENT MODULE  of the paper.
.PP
Once the composite WTN is produced, each correspondence set (CS) is
evaluated using the selected scoring function.  <A
HREF="rover/rover.htm#Section2.2"> Section 2.2. WTN VOTING SEARCH
MODULE: describes the voting process in detail.
There are three voting schemes described in the paper:
.RS
\*L By Word Frequency \*O

.RS
 To use word frequency as the scoring function, use	the options
'-m avgconf -a 1.0 -c 0.0'.  By setting -a to 1.0, the tradeoff
between  word occurrences and confidence scores, only the 
word occurrences are used.
.RE
.RE
.RS
\*L By Average Confidence Scores \*O

.RS
 The '-m avgconf' option make the voting function use
average confidence scores.
.RE
.RE
.RS
\*L By Word Maximum Confidence Scores \*O

.RS
 The '-m maxconf' option make the voting function use
the maximum confidence per word as the scoring metric.
.RE
.RE
.RS
.RE
REVISION HISTORY 
.RS
This is the initial release.
.RE
BUGS/COMMENTS 
.RS
Please contact Jon Fiscus at NIST with any bug reports or comments at
the email address 
\*Ljonathan.fiscus@nist.gov \*O or
by phone, (301)-975-3182.  Please include the version number of rover,
.RE