About UsApplicationsCoursesResearchConsultingBusinessTikiWikiContact us

  

Design and Implementation

of Speech Recognition Systems

This is a course in the actual development and construction of a speech recognition system.  That is, rather than merely studying the principles of speech recognition and how to use an existing speech recognition system, the project teams will develop their own speech recognition system components.

Projects:

  • Preliminary Projects

    • Unit test for other projects

    • Reading and Writing Common File Formats

    • Dictionary Processing

    • Text Processing

    • Recording Training Data

  • Training

    • Aligning Acoustic Data

    • Training Hidden Markov Models

    • Corrective Training

  • Acoustic Modeling

    • Phoneme-in-Context Modeling

    • Gaussian Mixture Distribution

  • Language Modeling

    • Statistical Language Models (n-Gram)

    • Back-off and Smoothing

  • Search

    • Frame Synchronous Search

    • Stack decoders

    • Generating n-Best Lists and Lattices

    • Multiple Passes and Rescoring

 

      
 

Copyright © 2005 James K. Baker