Electronic Theses and Dissertations

Identifier

366

Date

2011

Document Type

Thesis

Degree Name

Master of Science

Major

Electrical and Computer Engr

Concentration

Computer Engineering

Committee Chair

Mohammed Yeasin

Committee Member

Eddie Jacobs

Committee Member

Xiangen Hu

Abstract

Automatic recognition of emotion has a huge potential in several applications. In order to address such potential, researchers from diverse fields are collaborating together to build systems capable of recognizing human emotion. As a preliminary step towards such systems, many works are being done to automatically detect facial expressions. A technique generally termed as ``Model Based Technique'' has gained significant attention among the researchers for its utility in detecting facial expressions.However, methods currently used for evaluation of the performance of such systems have several flaws and inefficiencies. Due to these inefficient evaluation methods, it becomes difficult to compare among the systems from their literary descriptions. In this thesis, origins of such flaws are analyzed and efforts have been made to derive some solutions. As a part of this endeavor, a Three Level Evaluation (TLE) model has been proposed. In addition, some new and efficient assessment metrics have been suggested that can make faithful comparison of the systems.

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