Real-time American Sign Language (ASL) recognition system that uses OpenCV and MediaPipe for gesture tracking and recognition. The system captures live video from your webcam, identifies hand gestures, and displays the corresponding ASL letter on screen in real-time.
The goal of this project is to explore computer vision and machine learning techniques for human–computer interaction and assistive technologies. Install the following Python packages: numpy, matplotlib, seaborn, opencv-python, mediapipe, pandas, scikit-learn, joblib Open with Python 3.10! Newer Py does NOT support mediapipe !
Data Collection Run collect_data.py to capture gesture samples: Each time you press S, a new image is saved into the dataset (CSV or image folder). There are 20 000 entries in the CSV.
Model Training Run train_model.py to: Load the dataset, train machine learning models and automatically pick the best one and save it into model.pkl
Real-Time Gesture Detection Run detect_gestures.py to: Start webcam feed Show recognized ASL letters in real-time Press ESC to exit