🤟 Sign Detection using Python, OpenCV, and scikit-learn This project implements a real-time hand sign recognition system using a webcam feed. It leverages Mediapipe for hand landmark detection and scikit-learn (with a trained Random Forest Classifier) for gesture classification.
🔧 Tech Stack Python
OpenCV – for video capture and display
Mediapipe – for hand landmark detection
scikit-learn – for model training and prediction
NumPy – for numerical operations
SciPy – for preprocessing or support functions
Pickle – to load the trained model
📦 Features Detects hand landmarks in real-time from webcam input
Classifies custom hand signs (e.g., 'A', 'B', 'L')
Draws bounding boxes and labels on the detected hand
Lightweight and runs on CPU with TensorFlow Lite backend
🚀 How It Works Capture video feed using OpenCV.
Detect hand landmarks using Mediapipe.
Extract features by normalizing landmark positions.
Predict the gesture using a pre-trained RandomForestClassifier.
Display the predicted gesture on the video frame.