Silent Voice

• 

  Silent Voice

Inspiration

Silent Voice was inspired by the need to improve communication for individuals with hearing disabilities to communicate with people more friendly, aiming to create a more inclusive environment using technology.

What it does

Silent Voice is a web app that translates American Sign Language (ASL) into text with speech in real-time via a webcam, and includes an ASL dictionary, practice quizzes, and an admin panel to update signs.

How we built it

• Frontend: We built the user interface using Next.js, ensuring responsiveness and a clean design with Tailwind CSS.
• Backend: The backend was developed using FastAPI, a modern web framework that allowed us to build and serve the application efficiently.
• Machine Learning: We used MediaPipe for gesture detection and TensorFlow for training the machine learning model. The model itself is based on MobileNetV2, optimized for recognizing ASL gestures.
• Integration: The trained model was integrated with the FastAPI backend, enabling real-time processing and communication with the frontend. The frontend captures video input via the webcam, which is then processed by the model to output translated text.

Challenges we ran into

Key challenges included achieving high accuracy in gesture recognition and integrating the machine learning model seamlessly with the web application.

Accomplishments that we're proud of

We're proud of building a functional ASL translator with real-time capabilities and enriching it with features like a dictionary and practice tools.

What we learned

We learned about advanced machine learning techniques for gesture recognition, integration with FastAPI, and the importance of accessible technology.

What's next for Silent Voice

We plan to add support for more sign languages, improve translation accuracy, and explore speech-to-sign language features to broaden its impact.

Built With

• fastapi
• medianetv2
• mediapipe
• nextjs
• tailwindcss
• tensorflow