BRIDGE-AI: Sign Language Interpreter

Every great project starts with an idea. For me, that idea was born during the Digital Innovation Challenge at NYU Shanghai—a competition that pushed me to think creatively about how technology could solve real-world problems. Little did I know that this challenge would lead me to develop an assistive technology prototype, work with the deaf and hearing-impaired community, and create a mobile app to collect sign language gesture data.

The Spark: Finding the Right Problem to Solve

When I first entered the Digital Innovation Challenge, I wasn’t entirely sure what I wanted to work on. But one thing was clear—I wanted to create something with a real social impact. As I explored different problem spaces, I became fascinated by assistive technology and its potential to bridge communication gaps for people with disabilities.

After researching the challenges faced by the deaf and hearing-impaired community, I saw a major gap in AI-driven sign language recognition. Most existing models struggled with accuracy due to limited and non-diverse datasets. That’s when the idea hit me:

What if I could create an AI-powered system to recognize sign language gestures more accurately by collecting data directly from the community?

Prototyping: From Concept to Reality

Once I had my idea, the next step was bringing it to life. The Digital Innovation Challenge provided the perfect environment to experiment, get feedback, and iterate on my concept. I built an early prototype—a simple AI model trained to recognize a few sign language gestures using computer vision.

But very quickly, I ran into a major roadblock: data. AI models need high-quality datasets to function effectively, and existing sign language datasets were either too small or lacked diversity. I realized that if I wanted to create a truly accurate and inclusive model, I needed to collect data from real users.

Expanding the Vision: Taking It to the Community

The competition ended, but my project was far from over. Motivated by the potential impact of my idea, I decided to take it a step further. I reached out to Naga City’s PWD community, secured permission from city officials, and organized a workshop to introduce the deaf and hearing-impaired community to assistive technology—including my prototype.

To make data collection more accessible, I developed a Flutter-based mobile app, allowing participants to record sign language gestures on their own phones. This approach not only simplified the process but also ensured that the dataset reflected real-world signing variations, making the AI model more robust.

Lessons Learned: From Competition to Real-World Impact

What started as an idea in a university competition transformed into a community-driven project with real social impact. Along the way, I learned:

• Innovation starts with identifying real-world problems. Participating in the Digital Innovation Challenge taught me how to frame problems effectively and think beyond just building cool technology.
• AI is only as good as its data. The biggest challenge wasn’t the algorithm itself but ensuring the dataset was diverse, representative, and collected ethically.
• Collaboration is key. My biggest breakthrough came when I engaged with the deaf and hearing-impaired community. Their insights helped shape my project into something that could truly make a difference.

Looking Ahead

This journey is far from over. I’m now focused on improving the AI model, refining the dataset, and exploring ways to deploy this technology in real-world applications. What started as a simple idea in a competition has grown into something bigger—a mission to make assistive technology more inclusive and impactful.