Date of Award

6-1-2015

Document Type

Thesis

Publisher

Santa Clara : Santa Clara University

Department

Computer Engineering

First Advisor

Maria Pantoja

Abstract

A major part of having correct pronunciation when learning a new language is moving your lips in the correct way. This is a difficult thing to learn and to teach. One solution to this is software which tracks a student's lip movements and provides feedback. This paper describes how we have created a C++ library to accurately track lips in provided images. Further, this library attempts to use a CUDA-enabled GPU implementation to improve the algorithm's performance. It will fall back on a CPU implementation if such a GPU is not found. As a result, the lip tracking library runs on Windows, Linux, and OS X, as well as Android devices.

Share

COinS