The aim of the research work presented is to show a system that facilitates speech training for hearing impaired people. The system engineered combines both visual and acoustic speech data acquisition and analysis modules. The Active Shape Model method is used for extracting visual speech features from the shape and movement of the lips. The acoustic features extraction involves mel-cepstral analysis. Artificial Neural Networks are utilized as the classifier, feature vectors extracted combine both modalities of the human speech. Additional experiments with the degraded acoustic and/or visual information are carried out in order to test the system robustness against various distortions affecting the signals.

Authors: Dalka, Piotr; Kostek, Bozena
Affiliation: Gdansk University of Technology
AES Convention: 118 (May 2005) Paper Number: 6462
Publication Date: May 1, 2005 Import into BibTeX
Subject: Psychoacoustics, Perception, Listening Tests
Permalink: https://www.aes.org/e-lib/browse.cfm?elib=13178

Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!

This paper costs $33 for non-members and is free for AES members and E-Library subscribers.

Learn more about the AES E-Library

E-Library Location: (CD 118Papers)   /118/6462.pdf

Start a discussion about this paper!