Electronic Theses and Dissertations

Identifier

727

Date

2012

Document Type

Thesis

Degree Name

Master of Science

Major

Electrical and Computer Engr

Concentration

Computer Engineering

Committee Chair

Mohammed Yeasin

Committee Member

Bashir Morshed

Committee Member

Aaron L Robinson

Abstract

The aim of this study is to develop a multimodal co-analysis framework for continuous gesture recognition by exploiting prosodic and kinesics manifestation of natural communication. Using this framework, a co-analysis pattern between correlating components is obtained. The co-analysis pattern is clustered using K-means clustering to determine how well the pattern distinguishes the gestures. Features of the proposed approach that differentiate it from the other models are its less susceptibility to idiosyncrasies, its scalability, and simplicity. The experiment was performed on Multimodal Annotated Gesture Corpus (MAGEC) that we created for research on understanding non-verbal communication community, particularly the gestures.

Comments

Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.

Download

Share

COinS