Improving coarticulation performance of 3D avatar and gaze estimation using RGB webcam

Gu, Kuangxiao

Permalink

https://hdl.handle.net/2142/93004 Copy

Description

Title

Improving coarticulation performance of 3D avatar and gaze estimation using RGB webcam

Author(s)

Gu, Kuangxiao

Issue Date

2016-06-16

Director of Research (if dissertation) or Advisor (if thesis)

Huang, Thomas S.

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• 3D avatar
• Gaze estimation
• RGB camera

Abstract

This thesis explores two applications of computer vision in psychology-related studies: enhanced patient portal messages using 3D avatar and gaze estimation using a single RGB camera. The first application aims to help patients, especially those with poor health and low medical literacy, to understand messages delivered by patient portal systems by enhancing the messages with a 3D avatar. The avatar is built from real human face images and can deliver both semantic and emotive information, the latter of which is expected to help the patients to get a better, gist-level understanding of the portal messages. The second application aims to estimate eye gaze direction with an RGB camera. Preliminary results show the potential of the proposed method, although rigorous quantitative evaluation still needs to be done. While the proposed method cannot achieve the resolution and accuracy of commercial eye trackers, it is able to greatly reduce the cost since only one RGB camera is required.

Graduation Semester

2016-08

Type of Resource

text

Permalink

http://hdl.handle.net/2142/93004

Copyright and License Information

Copyright 2016 Kuangxiao Gu

Owning Collections