Implementation of a Real-Time, Interactive, Continuous Speech Recognition System.
Abstract
This thesis describes a speech recognition system designed and implemented to recognize continuous speech in a real time environment (after training). Several techniques were incorporated to characterize phonemes as vectors in space. Through the use of distance rules it was possible to characterize words by a phoneme representation, which could subsequently be used in word recognition. This approach to speech recognition offers several possibilities for future investigations such as varying the Minkowski distances and applying clustering techniques. The algorithm developed was modularized on a hierarchical basis and was user friendly. Additional keywords: software design; computer applications; computer programs; Fortran. (Author).
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1984
- Accession Number
- ADA151936
Entities
People
- K. R. Dixon
Organizations
- Air Force Institute of Technology