An open API service indexing awesome lists of open source software.

https://github.com/rahulkumar7189/american-sign-language-detection

Real-time American Sign Language detection system using CNN, OpenCV, MediaPipe, and PyTorch for computer vision-based sign language recognition
https://github.com/rahulkumar7189/american-sign-language-detection

ai american-sign-language asl cnn computer-vision deep-learning education kaggle-dataset machine-learning mediapipe opencv pytorch real-time-detection sign-language

Last synced: 2 months ago
JSON representation

Real-time American Sign Language detection system using CNN, OpenCV, MediaPipe, and PyTorch for computer vision-based sign language recognition

Awesome Lists containing this project

README

          

# American Sign Language Detection

## 📖 Project Overview
This project provides a comprehensive American Sign Language (ASL) detection system using computer vision and deep learning. The pipeline, training, and hand gesture recognition work accurately for all genders, including men, women, and non-binary individuals, ensuring inclusive and unbiased performance across diverse users.

## ✨ Features
- Real-time ASL gesture detection
- CNN-based deep learning model
- Hand landmark detection using MediaPipe
- Support for multiple ASL signs
- Training pipeline for custom datasets
- Gender-inclusive recognition optimized for all users

## 🚀 Usage

### Training the Model
```bash
python train.py
```

### Creating Custom Dataset
When collecting data using `create_dataset.py`, ensure diverse participant representation:
- Include hand gestures from male, female, and non-binary participants
- Capture various hand sizes, skin tones, and lighting conditions
- Maintain consistent gesture form across all demographics

```bash
python create_dataset.py
```

### Running Sign Detection
```bash
python sign_detector.py
```

### Evaluating the Model
```bash
python Evaluation.py
```

## 🧠 Model Architecture
The CNN model consists of:
- Convolutional layers for feature extraction
- Pooling layers for dimensionality reduction
- Fully connected layers for classification
- Dropout for regularization

## 📊 Dataset
The project supports various ASL datasets including:
- Custom datasets created using `create_dataset.py` (ensure gender-diverse data collection with male, female, and non-binary participants for robust model performance)
- Kaggle ASL datasets
- Hand landmark data from MediaPipe

**Note on Dataset Inclusivity**: For optimal performance across all users, training data should include hand gestures from participants of all genders, with particular attention to capturing male hand gestures alongside female and non-binary participants to prevent gender bias in recognition accuracy.

## 🎓 Applications
- **Education**: Teaching ASL to students
- **Accessibility**: Bridging communication gaps
- **Research**: Sign language recognition studies
- **Real-time Translation**: Converting ASL to text/speech

## 🤝 Contributing
Contributions are welcome! Please feel free to submit a Pull Request.

## 📝 License
This project is open source and available under the MIT License.

## 🙏 Acknowledgments
- MediaPipe for hand landmark detection
- PyTorch community for deep learning framework
- Kaggle for ASL datasets
- OpenCV for computer vision tools

## 📧 Contact
For questions or suggestions, please open an issue in this repository.

---
**Made with ❤️ for the ASL community**