Automatic Speech Segmentation

Description

Sebastian the researcher need a program to help with his research in find differences in reading of people between the silence reading part to the vocal reading part.  The researcher have database of voice of people reading and he need to divided the voice to words manually.

The aim of this project is to find an automatic method do find where word is ended and other word start.

Voice sample
Voice sample

Solution:

Using a BackPropegation Neuronal Network that trained on features that extracted from Praat and some samples that provided by the users, the student build a automatic method to divide the voice to words.

Students
1. Michael Kleinzit – [email protected]
2. Maria Uretskiy – [email protected]
Documentation (Hebrew)      Presentation A  Presentation B        Source Code            Praat

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Hananel Hazan , חננאל חזן , Activities, Academic Publications , Open Source Programs

Verified by MonsterInsights