Date of Award

Spring 2007

Project Type


Program or Major

Electrical Engineering

Degree Name

Master of Science

First Advisor

Andrew L Kun


Robust word boundary detection is essential for the efficient and accurate performance of an automatic speech recognition system. Although word boundary detection can achieve high accuracy in the presence of stationary noise with high values of SNR, its implementation becomes non-trivial in the presence of non-stationary noise and low SNR values. The purpose of this thesis is to compare and contrast the accuracy and robustness of various word boundary detection techniques and to introduce modifications to better their performance.