✋ Sign Language Recognition Using MNIST Hand Gesture Dataset 🤟📊

Sign-Language-Recognition-Using-MNIST-Hand-Gesture-Dataset is a computer vision and deep learning project that recognizes American Sign Language (ASL) hand gestures using the MNIST Hand Gesture Dataset. The project demonstrates how Convolutional Neural Networks (CNNs) can be applied to empower communication accessibility for the deaf and hard-of-hearing community.

✨ Key Features

✋ Gesture Recognition: Classify hand signs into ASL alphabet letters (A–Z, excluding J and Z due to motion)

🖼️ Dataset: Trained on the MNIST Hand Gesture Dataset (image-based labeled hand signs)

🧹 Preprocessing: Grayscale normalization, resizing, and data augmentation for robustness

🧠 Deep CNN Models: Custom CNNs and pretrained models (VGG16, ResNet, MobileNet)

📊 Performance Metrics: Accuracy, Precision, Recall, F1-score, Confusion Matrix

📈 Visualization: Training/validation curves, heatmaps, and Grad-CAM to show model focus

🌐 Real-Time Demo: Deployable web app for live gesture recognition (via webcam)

🧰 Tech Stack

Programming: Python 🐍

Deep Learning: TensorFlow / Keras or PyTorch

Libraries: NumPy, Pandas, OpenCV, Matplotlib, Seaborn, Scikit-learn

Deployment (Optional): Flask, Streamlit, FastAPI

📁 Project Structure 📁 dataset/ # MNIST Hand Gesture Dataset 📁 preprocessing/ # Data cleaning & augmentation scripts 📁 models/ # CNN and pretrained architectures 📁 notebooks/ # Jupyter notebooks for training & evaluation 📁 results/ # Metrics, plots & Grad-CAM visualizations 📁 app/ # Web app for real-time sign recognition

🚀 Getting Started git clone https://github.com/yourusername/Sign-Language-Recognition-Using-MNIST-Hand-Gesture-Dataset.git cd Sign-Language-Recognition-Using-MNIST-Hand-Gesture-Dataset pip install -r requirements.txt jupyter notebook

📌 Use Cases

🧏 Accessibility: Helps bridge communication gaps for the deaf community

📱 Applications: Integration into mobile apps for real-time sign recognition

🏫 Education: Assisting learners in practicing ASL alphabets

🤖 Research: Benchmark for applying CNNs in gesture and pattern recognition

🤝 Contributing

Contributions are welcome! You can add more architectures, improve dataset preprocessing, or expand to full ASL word recognition.

📜 License

MIT License – Free to use for research, learning, and open-source development.

⭐ Support

If you find this project valuable, please give it a star ⭐ to support open-source work in AI for accessibility.