EyeSpeak

Project Story

Inspiration

We were inspired by a UCSD water polo player who was involved in a scooter accident and was left temporarily paralyzed and unable to speak. Seeing how isolating that experience was made us realize how difficult it is for patients to communicate even basic needs, pushing us to build this.

What We Learned

As we researched the problem, we realized the number of people who could benefit from this is much larger than we initially expected. Stroke patients, ICU patients, and others with temporary loss of speech or motor function all face similar challenges. What started as a niche idea quickly expanded into something with a much broader impact.

How We Built It

We built EyeSpeak using pure HTML, vanilla JavaScript, and advanced CSS to keep it lightweight and accessible through any browser. For eye tracking, we used WebGazer.js, a client-side library developed by Brown University. We also integrated the Web Speech API to provide real-time text-to-speech, allowing patients to communicate their needs instantly.

Challenges

The biggest challenge was implementing reliable eye tracking, especially while working with limited hardware like standard webcams. Accuracy and calibration were difficult to manage, and small inconsistencies could impact usability. Designing an interface that remained simple and usable despite these limitations was a key part of our process.

Built With

• html
• javascript