
An interactive Computer Vision pipeline that transforms your webcam into a gesture-controlled interface using Googleβs MediaPipe and Python.
π΄ Live Preview
π Key Features
- β‘ Real-time Detection: High-speed hand landmark tracking (21 unique points).
- π Colab Native: Custom JavaScript bridge to access local webcams within virtual environments.
- π§ Logic-Based Classification: Robust rule-based engine to identify:
-
π Open Palm |
βοΈ Peace |
β Fist |
π Thumbs Up |
βοΈ Pointing |
- π¨ Augmented Overlay: Dynamic rendering of the hand skeleton and gesture labels.
π οΈ How It Works
The Pipeline
- Capture: JavaScript initiates a webcam stream and captures a frame.
- Processing: OpenCV converts images to RGB; MediaPipe maps 21 3D landmarks.
- Inference: * Fingers: Compares
TIP vs PIP y-coordinates to determine βExtendedβ vs βCurledβ.
- Thumb: Analyzes horizontal orientation relative to the palm base.
- Feedback: The system annotates the image and renders it instantly.
π Getting Started
1. Launch Environment
Open this project in Google Colab to ensure the webcam JavaScript bridge executes correctly.
2. Install Dependencies
```bash
pip install mediapipe opencv-python numpy