Silent Talk

Real-Time Sign Language Translation & Accessibility Platform

Table of Contents

1. About The Project
   • Built With
2. System Architecture
3. Getting Started
   • Prerequisites
   • Installation
4. Status & Disclaimer
5. License
6. Contact

About The Project

Silent Talk is a computer vision accessibility tool designed to bridge the communication gap for the deaf and hard-of-hearing community. By translating American Sign Language (ASL) gestures into text in real-time, it enables seamless two-way communication.

This project was selected as a Top 30 Semifinalist (out of 300+ teams) in the Samsung Solve for Tomorrow 2024 competition.

• Metric: Trained on a custom dataset of ~50,000 images.
• Performance: Achieved >90% accuracy on validation samples during competition trials.
• Optimization: Engineered to run on standard CPU hardware without requiring GPU acceleration.

(back to top)

Built With

• 
• 
• 
• 
• 

(back to top)

System Architecture

The system is designed as a monolithic prototype optimized for the rapid iteration constraints of a hackathon environment.

1. Vision Engine (Backend)

Located in /backend, the core logic utilizes a custom lightweight pipeline:

• Hand Tracking: MediaPipe extracts skeletal landmarks (21 points per hand) in real-time.
• Gesture Classification: A trained Scikit-Learn classifier maps the vector geometry to ASL alphabets.
• Speech Synthesis: Recognized text streams are converted into audio output.

2. User Interface (Frontend)

Located in /frontend, the client is built with React:

• Real-Time Feedback: Visualizes the translation stream with low latency.
• Accessibility: High-contrast UI designed for diverse user visibility needs.

(back to top)

Getting Started

To run this research prototype locally, ensure you have Python 3.9+ and Node.js installed.

Prerequisites

• Python 3.9+
• Node.js & npm

Installation

1. Clone the repository

git clone https://github.com/alexandr-tk/silent-talk.git

2. Setup Backend (Python)

cd silent-talk/backend
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt
python app.py

3. Setup Frontend (React)

cd ../frontend
npm install
npm run dev

(back to top)

Status & Disclaimer

Status: Archived (Proof of Concept)

This codebase represents the competition snapshot of Silent Talk submitted for the Samsung Solve for Tomorrow semifinals. It was optimized for speed of development and proof-of-concept validation.

• Code Quality: Experimental / Prototype Grade.
• Maintenance: This repository is archived and is not currently maintained.

(back to top)

License

Distributed under the MIT License. See LICENSE for more information.

(back to top)

Contact

Alex Tkachyov - Lead Developer - LinkedIn

Project Link: https://github.com/alexandr-tk/silent-talk

(back to top)