Hand Tracking Cursor

Control your computer's cursor using hand gestures via your webcam.

Features

• Hand tracking - Detects and visualizes all 21 hand landmarks with skeleton overlay
• Cursor control - Move your cursor by moving your hand in front of the camera
• Pinch-to-click - Pinch your index finger and thumb together to click
• Dead zone - Small hand movements are ignored to reduce cursor jitter
• Fingertip labels - Each fingertip is labeled on the camera feed (thumb, index, middle, ring, pinky)

How It Works

• The palm center is calculated by averaging 5 palm landmarks (wrist + base of each finger)
• Palm position is mapped from camera coordinates to screen coordinates to move the cursor
• A dead zone threshold prevents jittery cursor movement when your hand is still
• Pinch detection measures the distance between the index fingertip and thumb tip — when close enough, a click is triggered

Requirements

• Python 3.10+
• A webcam

Setup

1. Create a virtual environment and activate it:

   python -m venv venv
   source venv/bin/activate

2. Install dependencies:

   pip install opencv-python mediapipe pyautogui

3. Download the MediaPipe hand landmarker model and place it in the project root:

   • hand_landmarker.task

4. On macOS, grant Accessibility permissions to your terminal/Python so pyautogui can control the cursor:

   • System Settings > Privacy & Security > Accessibility

Usage

python main.py

• Move your hand to control the cursor
• Pinch index finger + thumb to click
• Press q to quit

Configuration

You can tweak these values at the top of main.py:

Variable	Default	Description
cv2.VideoCapture(0)	0	Camera index (try 1 or 2 if default doesn't work)
DEAD_ZONE	10	Minimum cursor movement threshold (higher = more stable, less responsive)
PINCH_THRESHOLD	30	Max distance between index and thumb to trigger a click (higher = easier to click)