This project focuses on real-time Sign Language Translation (SLT) using computer vision and deep learning. It uses keypoint extraction to recognize gestures and translate them into text or speech.
- Keypoint Extraction: Uses MediaPipe (or similar) to capture hand and body movements.
- Data Collection: Tools to record and append new sign language keypoints to the dataset.
- Model Training: Scripts to train a neural network on the collected keypoint data.
- Live Prediction: A real-time interface to predict signs using a webcam.
data/: Directory containing the collected keypoint datasets.utils2.py: Utility functions for processing landmarks and drawing on the screen.append_new_keypoints.py: Script to add new gesture data to your training set.train2.py: The main script to train the gesture recognition model.live_predict2.py: The script for running real-time inference via webcam.test_camera.py: A simple script to verify your camera setup.