Abstract:

Eye based interface systems for computers have existed for several years, but have not come into widespread use due to the high cost of the associated technology and devices. Recent developments in the area of high grade consumer cameras with small form factors have created devices which allow for low cost gaze tracking research to be conducted. Previous design work done in this field has often neglected to provide feedback to the user, thereby limiting the potential uses of those systems. The designed Gaze Tracking System attains similar tracking results compared to previously designed systems, while also providing visual feedback to the user in the form of a computer cursor that follows the user’s gaze. Results have doubled the previous best tracking accuracy to .5 , as well as achieving a 12% speed increase over conventional mouse based systems. Additionally, a Head Mounted Display has been integrated into the system, thereby allowing full hands free use and mobility. Continuing advances in this field, such as those shown in this research, will rapidly lead to the technology being viable for real world use.

What the project does:

Currently, we have designed a system for full active eye tracking. An experimental frame holds a camera close to a user's eye, allowing us to continually capture images of the eye. Image processing is used to determine the location of the user's gaze. A screen capture from the real-time tracking system is shown below.

The colors overlaid on the image are directly generated by the tracking system. The blue region represents the location of the user's pupil as tracked by the system. From this blue region, a red best fit circle is determined. The green dot indicates the center of the circle, which is the calculated position of the user's eye. This position is then mapped to a location on a screen, so the computer cursor actively follows wherever the user is looking on the screen.

More details can be found within our project proposal.

Final Presentation and Demo

05.13.2008

Presentation: Our final presentation deliverable was due on May 1st. This presentation went well and we received compliments from both the professors and the students who watched the presentation.

Demonstration: Last week, on May 8th, we gave our final demonstration to the professors in accordance with the ABET requirements. Recent trials of the software had been going so well for us that Murphy’s Law dictated that this demo couldn’t go well, so inevitably, it didn’t. Nothing worked when we tried it, the system just crashed. After trying a number of quick solutions we were unable to find the source of the error. We eventually discovered the problem was related to cross-compiler issues between Visual Studio 2008 and Visual Studio 2005. We were able to fix the issue by releasing the user interface, designed in VS08, as a stand-alone application that launches the image processing application, created in VS05. This immediately resolved the problem we were having and the system went back to functioning as it had previously. We called back the professors and successfully demonstrated the system. After this experience, we will be properly packaging the installer for the application so that these issues do not appear again.

   - Breanna Heidenburg

Engineering Advisory Board Presentation

04.25.2008

Today, we gave our presentation to electrical engineering advisory board members. We were able to use the same poster we used at the student expo for this presentation. Overall, this presentation went very well and the advisory board asked a lot of good questions. It was a good experience to be able to speak to engineering managers about a product or project and be able to get away from the technical side a bit to see what kind of questions we'd really be asked in the real world. Each of our bosses were at our presentation; they all seemed reasonably impressed and declared that we did a good job.

   - Breanna Heidenburg

System Functional

04.07.2008

We recently integrated all of the system components. Our system successfully tracks a user's eye and positions a cursor on the display. We demonstrated this functionality at the Bradley University Student Exposition on Friday, April 4th.Shown here is the poster that was used at the expo:

Along with the expo, we now have a set of experimental results. Our research expected to determine the performance of the system on two separate levels. The first topic was recognition and cursor placement accuracy. In order to test this, we have created a method for determining the error in the system. This method generates a plot for pixel error. This pixel error is shown here:

This error plot shows that the pixel error for our 4th order mapping is most prevalent in the corners of the screen. Along with measuring the error in our recognition and calibration systems, we were interested in measuring the reaction time of the system. It was our initial assumption that by using an eye based system we would be able to speed up Human-Computer Interaction. Our initial results confirmed our assumptions, and are shown here:

  - Michael Lenisa