Synchrony-Based Processing for Robust Speech Recognition

Fall 2009

Penelope Tsatsoulis
Richard Stern
Project description

The student's project will be to reevaluate a synchrony-based processing for robust automatic speech recognition algorithm developed by Chanwoo Kim, Yu-Hsiang Chiu, and Richard M. Stern. This algorithm utilizes the synchrony elements of an input signal in order to improve speech recognition accuracy.

The first stage of the student's work will be to evaluate the existing methods used in the project. This will include running experiments and verifying the code and processes used. At the end of the process the student should be able to completely confirm and support the original method, or provide an alternative process that improves results.

If an alternate process is required the student will implement the working phases of the previous model along with necessary modifications. The student will accomplish this by researching techniques and running experiments to verify results.

Return to project list