Scribd
Upload
391 views

Eye Tracking Proposal

This proposal requests funding from the Computer Science department to develop an affordable eye tracking library and drivers. The students will experiment with head-mounted and stationary cameras. They aim to create software that analyzes eye movements to provide input and for research applications. Over the course of 9 months, they will design cameras, develop algorithms to measure gaze, create an eye tracking library, develop input drivers, add prediction capabilities, and build applications like a gaze analyzer. The students believe their experience qualifies them to complete the project successfully. They request funding for cameras to modify for the research.

Uploaded by

Marcela Reyna
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
391 views

Eye Tracking Proposal

This proposal requests funding from the Computer Science department to develop an affordable eye tracking library and drivers. The students will experiment with head-mounted and stationary cameras. They aim to create software that analyzes eye movements to provide input and for research applications. Over the course of 9 months, they will design cameras, develop algorithms to measure gaze, create an eye tracking library, develop input drivers, add prediction capabilities, and build applications like a gaze analyzer. The students believe their experience qualifies them to complete the project successfully. They request funding for cameras to modify for the research.

Uploaded by

Marcela Reyna
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Memo

Date: July 31, 2006


To: Dr. Chen, Computer Science Department Chair
From: Arm Surwaranaratana
Marcela Reyna
Javier Villarreal
Subject: Proposal for student research on eye tracking hardware and software

Purpose
The purpose of our proposal is asking the Department of Computer Science to fund a
student research in which we will develop an eye tracking library and drivers.

Summary
This proposal is asking the Department of Computer Science at the University of Texas
Pan American to fund our student research on eye-tracking software and hardware and its
applications. We will experiment with cameras using two separate placement positions.
In our first position, one camera is mounted on the head and records the eye and the other
camera points to the user’s view which is then used as a frame of reference. In the second
position, only one camera is used which is stationary in front of the monitor pointing
towards the user. Our aim is that the user will benefit from our inexpensive software and
hardware suite by applying it to a variety of different research topics.

The research currently available on eye-tracking devices and software has focused on
learning how people respond to certain web, magazine, and advertisement layouts. An
example involving web layouts is discovering whether the user carefully looks at certain
sections of a webpage or merely glances over it. Using eye tracking software, researchers
have monitored patterns people follow when reading information online. The beginning
left side of the webpage is usually carefully read, while the right side of the page is then
skipped over. Studies have also shown that smaller fonts tend to keep the user’s attention
longer while larger fonts are overlooked.

Since computer vision solutions are typically expensive, we would like to create a
software and hardware suite which is affordable for the general public. We want to
provide smaller research groups an inexpensive solution. As part of this suite, we want to
develop a learning algorithm and extensive tools in order to provide even more
complexity and flexibility to our solution. These tools include a library, input drivers, and
utilities which will make use of the library and “gaze” analysis software.

We have Dr. Tsai, a Computer Science professor at the University of Texas Pan
American, as a contact whose expertise is computer vision. With his help, and the
assistance of other students we would like to begin our research August 24, 2006 and
continue until the end of the spring 2007 semester.
Introduction
At the University of Texas Pan American students are encouraged to participate in
undergraduate research. Unfortunately, most students are unaware that funds are
available. There are two reasons for our research; the first is to develop an affordable eye
tracking solution for input and research aid. The second is to encourage others to propose
their own ideas.

Eye tracking research and development has taken place in the field of Human Computer
Interaction (HCI). But many of the current solutions involve large ‘bulky’ head mounts
that are intrusive and uncomfortable. The current solutions are expensive and price
changes depending on equipment used and complexity of the software. Development is
lacking for every day use of eye tracking as primary input. Our goal is to improve current
eye tracking algorithms for both head mounted cameras and stationary camera solutions.
Our final development will be an open source set of user input drivers and software to
utilize eye tracking technology.

As mentioned previously, there has already been research in this field, but there are no
practical input solutions. There have been several algorithms available for eye-tracking
that we may freely build on. Our status throughout the year will be documented on our
website which is open for anyone to view or use our work.

Proposed Procedure
The following tasks will be done: first we will work with camera placement then develop
an algorithm to measure a user’s point of gaze and finally create an eye tracking library
along with helper applications and prediction algorithms.

1. Design camera placement for data acquisition


2. Develop technique to measure the user’s point of gaze
3. Combine development to create unified eye tracking library
4. Utilize eye tracking library to create input drivers
5. Create prediction algorithms
6. Create helper applications

Task 1: Design camera placement for data acquisition


The aim for the research will require us to use both the head mounted and stationary
camera placements which will allow us to gain maximum data results. We want to have a
single stationary eye tracking camera as our end product, but since head mounted
cameras yield simple data sets, we will be using that placement to develop the
preliminary eye tracking techniques. After working with camera placement, we will also
include infrared lighting to illuminate the eyes allowing us to easily identify it.

Task 2: Develop technique to measure the user’s point of gaze


There already exists a technique to extract the location of the pupil which is called the
Starburst algorithm and is implemented in many applications; however, the algorithm can
be inaccurate when the pupil is not completely in view. We intend to build on the
algorithm and/or derive our own algorithm to do pupil center estimation and correct the
problems the Starburst algorithm has so that there will be accurate results even if the user
blinks or blocks the pupil in any way. Infrared reflectors will help in measuring the user’s
point of gaze and will be discretely worn by the user.

Task 3: Combine all development to create unified eye tracking library


Many applications are developed from eye tracking technology. To make programming
more efficient for application programmers, we will create a unified eye tracking library
which will provide an eye tracking hardware abstraction layer. Such a library would
allow independent development of eye tracking technology and applications.

Task 4: Utilize eye tracking library to create input drivers


Utilizing the eye tracking library, we will create input device drivers to simulate the use
of a mouse. We will also conduct research on creating new input methods with eye
tracking concepts.

Task 5: Create prediction algorithms


Our research is not restricted to development of an input device to replace a mouse or
keyboard; we would also develop prediction algorithms to aid the user in everyday
computer interaction. Such prediction algorithms will aid the end user by having
automatic scrolling, automatic menu pop ups, predicting if the user wants to switch to the
background window, and more human-computer interaction which we will research
further on once we reach this task.

Task 6: Create applications


For the eye tracking technology to be useful, it would need a set of applied software. A
gaze analysis application would be developed to allow the study of user interfaces in the
field of psychology, communication, and other departments at the University of Texas
Pan American. Additional applications include an eye driven window manager and eye
gesture based user interactions.

Schedule
Activity Start Date Finish Date
Design camera placement 1 Sept. 06 22 Sept. 06
for data acquisition
Develop technique to 25 Sept. 06 1 Nov. 06
measure the user’s point of
gaze

Combine all development to 2 Nov. 06 2 Dec. 06


create unified eye tracking
library
Utilize eye tracking library Dec 3. 06 7 Jan. 07
to create input drivers
Create prediction 8 Jan. 07 8. Feb 07
algorithms
Create applications 9 Feb. 07 9 Apr. 07
Qualifications and Experience
Our group members each possess a wide range of different skills which when combined
makes us effective and corporative together.

Arm Suwaranaratana is the project leader and has a strong interest in HCI (human
computer interaction). Arm brings a robust background in HCI research and image
processing to the project. He previously has developed hardware drivers for the Winbond
W83627THF sensor board for the Windows XP operating system and a kernel patch for
ARP spoofing vulnerability for Linux kernel with proof of concept. His extensive
programming experience will help drive software development.

Marcela Reyna is a lab consultant with the University of Texas Pan American where she
has gained experience in working with groups and helping others. As a result, Marcela
knows how to keep everyone focused and promote productivity. She is a hard worker
who will put in the extra effort when learning new concepts. Her expertise in
mathematics will be invaluable to the algorithm development of the project.

Javier Villarreal is currently employed with Smartcom Inc as Computer Technical


Support. Javier understands how to communicate well with people because of his skill in
dealing with customers. Javier has experience with UI design and development that will
be useful in creating a friendly and intuitive interface for the software solution. His
analytical nature and attention to detail will also help maintain and prevent errors during
the lifetime of the project.

Budget
The budget for our research is shown below. The following equipment is not available
from other departments and we will be modifying them.

Number of Items Item Amount


6 Fire-I firewire monochrome cameras 640x480 at 30 $150
fps
2 Fire-I firewire color cameras 640x480 at 30 fps $99
2 DMK 31AF03 firewire camera mono 1024x768 at 30 $700
fps
2 Firewire Adapter $150

Total Cost: $2650

You might also like