Real-Time Sign Language Translation Using LSTM and MediaPipe ✨

Overview

The Sign Language Translator is a real-time system that captures and translates sign language gestures into text. Utilizing computer vision and deep learning, it enables seamless communication between sign language users and non-signers by identifying hand, face, and body movements. The system processes video input, detects keypoints using MediaPipe Holistic, and predicts gestures with an LSTM-based neural network trained on sign language datasets.

Features

Real-time Gesture Recognition

• Detects and translates sign language gestures instantly.

Multi-Body Landmark Tracking

• Recognizes hand, face, and body movements using MediaPipe Holistic.

Deep Learning-Based Prediction

• Utilizes LSTM neural networks for accurate classification of gestures.

Technologies:

Programming & Libraries

• Python: Main programming language for data processing and deep learning.
• OpenCV: Handles real-time video processing and UI display.
• MediaPipe: Detects and tracks body keypoints for gesture analysis.
• NumPy: Efficient numerical computations and data manipulation.

Machine Learning & Deep Learning

• TensorFlow & Keras: Used to build and train the LSTM-based neural network.
• Scikit-learn: Provides utilities for model evaluation and data processing.

Hardware & Environment

• Webcam: Captures video input for real-time processing.
• Jupyter Notebook: Used for developing and testing the model.

Screenshot 💻

Author

Developed by 3 Members on 2024

If you appreciate our work, consider adding this project to your favorites on GitHub. DM me if you’d like to collaborate with us.