Eye-Tracking Game Control for Accessibility

ISEF Category: Technology Enhances the Arts

Subcategory: Games  ·  Difficulty: Intermediate  ·  Setup: Home Setup  ·  Time: 1 to 2 Months

The Hook

A webcam can do more than take your picture. With the right code, it can turn your gaze into a game controller. That matters for players who cannot use a mouse well, because aim can decide whether a game feels fun or impossible. Your project can measure whether eye tracking actually helps.

What Is It?

This project tests an accessibility game control system. Instead of moving a mouse to aim, the player looks at targets on screen, and the webcam estimates where the eyes are pointing. WebGazer.js is one tool that can estimate gaze from webcam video, and a Kalman filter is a math tool that smooths noisy tracking data so the cursor does not jump around.

Think of it like trying to follow a finger in a shaky video. Raw tracking data wiggles a lot. Filtering helps the system guess the true target more steadily. Your job is to compare that eye-driven control with mouse control and see how well each one works for accuracy, speed, and fatigue.

You are not just building a game. You are testing an interface. That means you can ask whether the design helps players hit targets faster, miss less often, or feel less tired after repeated play.

Why This Is a Good Topic

This is a strong science fair topic because you can measure it with clear numbers. You can track hit rate, reaction time, target error, and self-reported fatigue, then compare eye control against mouse control. The project connects to accessibility design, assistive technology, and human-computer interaction. You can learn how to plan a fair user test, clean noisy data, and analyze whether one input method really performs better.

Research Questions

• How does eye-tracking control affect hit rate compared with mouse control for a simple target shooter?
• What is the effect of Kalman filtering on aim stability and missed shots in webcam-based gaze control?
• Does target size change the performance gap between eye tracking and mouse input?
• To what extent does calibration quality affect gaze accuracy across repeated trials?
• Which control mode produces lower reported fatigue after the same play session?
• How does cursor smoothing change response time and error rate in users with limited motor function?

Basic Materials

• Laptop or desktop computer with webcam.
• Web browser that supports WebGazer.js.
• Simple target-shooter game built in JavaScript or a similar web tool.
• External mouse.
• Timer or built-in logging script.
• Spreadsheet software for data entry and charts.
• Consent form and short fatigue survey.
• Quiet room with stable lighting.

Advanced Materials

• Computer with a high-quality webcam.
• Browser-based gaze-tracking prototype using WebGazer.js.
• Python or R for statistical analysis.
• ImageJ or another free image analysis tool for screen capture checks.
• Screen recording software for trial review.
• External mouse with adjustable sensitivity settings.
• Optional head-position guide or chin rest for repeatability.
• Survey tool for structured usability and fatigue ratings.

Software & Tools

• WebGazer.js: Estimates gaze position from webcam video for a browser-based control prototype.
• JavaScript: Lets you build the game logic, logging, and control comparisons in one web app.
• Python: Helps you clean data, compute summary statistics, and make comparison plots.
• Google Sheets: Lets you organize trial results and calculate simple averages and graphs.
• ImageJ: Helps inspect screenshots or cursor tracks if you want to compare aim spread visually.

Experiment Steps

1. Define the player task, the control modes, and the exact outcomes you will measure.
2. Build a simple game that logs hits, misses, reaction time, and cursor error in the same way for both input methods.
3. Set up calibration and filtering rules so the eye-tracking mode has a consistent starting point.
4. Plan a fair comparison with the same target layout, same score rules, and the same number of trials for each mode.
5. Choose data summaries that capture both performance and fatigue, then decide how you will compare participants or trial blocks.
6. Prepare controls for lighting, screen distance, and break timing so outside factors do not dominate the results.

Common Pitfalls

• Testing in changing light, which makes webcam gaze estimates drift from one session to the next.
• Comparing eye tracking and mouse control with different target layouts, which makes the results unfair.
• Letting calibration fail silently, which creates bad cursor placement and fake misses.
• Using only one score metric, which hides the tradeoff between speed, accuracy, and fatigue.
• Forgetting to separate practice trials from test trials, which lets learning effects blur the comparison.

What Makes This Competitive

A strong version of this project goes beyond a simple side-by-side comparison. You can test whether filtering, calibration quality, or target design changes the gap between eye control and mouse control. You can also use stronger analysis, like comparing error distributions or testing whether fatigue rises across repeated rounds. That kind of design shows that you understand both the user experience and the data behind it.

Project Variations

• Compare eye tracking with mouse control for a rhythm game instead of a shooter, then measure timing error and fatigue.
• Test how different cursor smoothing methods change gaze accuracy for users with limited motor function.
• Compare gaze control on a laptop webcam versus an external webcam to see whether camera quality changes performance.

Learn More

• WebGazer.js documentation: Read the project notes and setup guidance on the official GitHub repository.
• NIH PubMed: Search for review articles on eye tracking, assistive technology, and human-computer interaction.
• MIT OpenCourseWare, Introduction to Algorithms: Use lecture materials to understand filtering, estimation, and signal smoothing ideas.
• NASA Open Science resources: Browse free data and visualization guides to practice clean analysis and charting.
• Human-Computer Interaction journals on PubMed and university libraries: Search for studies on accessibility interfaces and gaze-based input.