Flask App for Sign Language Recognition

This Flask app utilizes a trained model for sign language recognition using MediaPipe and TensorFlow/Keras.

Setup Instructions

1. Clone the repository:

   git clone <repo-url>
   cd <repo-directory>

2. Install dependencies:

   pip install -r requirements.txt

3. Run the Flask app:

   python app.py

4. Access the app in your browser at http://localhost:3000.

Usage

1. Visit the app in your browser.
2. Upload a video file.
3. Wait for the app to process the video and display the predicted actions.
4. Interpret the predictions based on the recognized actions.

Note: The model currently recognizes the following 6 classes of actions:

• Drink
• Eat
• Goodbye
• Hello
• Help
• How are you

Dependencies

• Flask
• NumPy
• TensorFlow
• Matplotlib
• MediaPipe
• OpenCV (opencv-python-headless)

File Structure

• app.py: Contains the Flask application code.
• requirements.txt: Lists all Python dependencies required for the app.
• model_lstm_6_classes_0.98.h5: Pre-trained LSTM model for action recognition.
• index.html: HTML template for the web interface.