Facial expression recognition in Filipino sign language

Date of Publication

2017

Document Type

Master's Thesis

Degree Name

Master of Science in Computer Science

College

College of Computer Studies

Department/Unit

Computer Science

Thesis Adviser

Clement Y. Ong

Defense Panel Chair

Conrado D. Ruiz, Jr.

Defense Panel Member

Jocelyn W. Cu
Merlin Teodosia C. Suarez

Abstract/Summary

Filipino Sign Language (FSL) is a mode of communication for the Deaf in the Philippines. Though it has two components, namely manual and non-manual signals, most research works focus on manual signals only. However, non-manual signals play a significant role in Sign Language Recognition (SLR) as it can be mixed freely with manual signals, often changing the meaning of the signs. Internationally, there have been numerous researches regarding non-manual signals. However, most of these focused on the semantic and lexical functions only.

This study focused on recognizing facial expressions in FSL that convey Types of Sentences (i.e. question, statement, and exclamation), Degrees of Adjectives (i.e. absence, presence, and high presence), and Emotions (i.e happy, fear/surprise, sad, and disgust/anger). The data were collected using Microsoft Kinect for Windows 2.0. After the data annotation, co-occurrences of the different categories were discovered. Thus, individual experiments were conducted for each category.

Based from the results of extensive experiments, head rotation angles and Animation Units of different facial features from peak facial expressions are not enough to represent each category of the facial expressions in FSL. Adding classes from other categories as features and reducing features through Genetic Algorithm generally improved the performances significantly. Based from the observations of the raw data, some important features are not represented. Thus, further research on the data representation is necessary to improve the performances.

Abstract Format

html

Language

English

Format

Electronic

Accession Number

CDTG007276

Shelf Location

Archives, The Learning Commons, 12F Henry Sy Sr. Hall

Physical Description

1 disc ; 4 3/4 inches

Keywords

Sign language; Philippine Sign Language; Facial expression--Philippines

Upload Full Text

wf_no

This document is currently not available here.

Share

COinS