Converting Finite State Grammar Representations from XML to SLS Format for Use in Multiple Speech Recognition Systems

Presenter Information

Tung Pham, Minnesota State University, MankatoFollow

Location

CSU 253

Start Date

25-4-2006 10:30 AM

End Date

25-4-2006 11:30 AM

Student's Major

Computer Information Science

Student's College

Science, Engineering and Technology

Mentor's Name

Colin Wightman

Mentor's Department

Computer Information Science

Mentor's College

Science, Engineering and Technology

Second Mentor's Name

Rebecca Bates

Second Mentor's Department

Computer Information Science

Second Mentor's College

Science, Engineering and Technology

Description

Rapid development of an automatic speech recognition system is constrained by the availability of software tools and format constraints for the statistical models used in the process. Development of models is often easier in one environment than another but the output may not be usable in all recognition systems. This project involves developing software tools to convert a file in XML (Extensible Markup Language) to a standard lattice format (SLS) so that it can be used as input to the HTK Hidden Markov Model Toolkit, an open-source speech recognition system. Information about the two formats and the process of conversion will be presented.