SignBridge: ASL to Text Translation

AI-powered American Sign Language (ASL) to text translation system that helps bridge communication gaps between the deaf/hard-of-hearing community and others.

Features

• Real-time ASL hand sign recognition
• Support for 24 ASL letters (A-Y, excluding J and Z)
• High accuracy (99.80% on test set)
• User-friendly web interface
• Confidence score display for predictions

Tech Stack

• Frontend: Streamlit
• Backend: Python, TensorFlow/Keras
• Data Processing: OpenCV, NumPy, Pandas
• Model: Convolutional Neural Network (CNN)

Installation

1. Clone the repository:

git clone [your-repository-url]
cd signvision

2. Install dependencies:

pip install -r requirements.txt

Usage

Run the application:

streamlit run app.py

The application will open in your default web browser. Upload an image of an ASL hand sign, and the model will predict the corresponding letter.

Model Architecture

• Input Layer: 28x28x1 (grayscale images)
• Multiple Convolutional Layers with BatchNormalization and MaxPooling
• Dense Layers with Dropout for regularization
• Output Layer: 24 classes (ASL letters A-Y, excluding J and Z)

Dataset

The model is trained on the Sign Language MNIST dataset from Kaggle, which contains 27,455 training and 7,172 test images.

Performance

• Test Accuracy: 99.80%
• Support for real-time processing
• Robust to various hand sizes and positions

Future Enhancements

• Support for complete ASL alphabet including J and Z
• Gesture recognition capabilities
• Sentence-level translation
• Mobile application development
• Enhanced UI/UX features

Team

• [Team Member 1] - [Role]
• [Team Member 2] - [Role]
• [Team Member 3] - [Role]
• [Team Member 4] - [Role]

License

[Choose an appropriate license]

Acknowledgments

• Sign Language MNIST dataset from Kaggle
• [Add any other acknowledgments]