Hand Gesture Control System

A real-time hand gesture recognition application that allows you to control your computer (scrolling and media playback [TODO]) using simple hand movements captured by your webcam.

Built with MediaPipe for hand detection and OpenCV for camera handling.

Features

• Scroll Up and Scroll Down using index and middle finger gestures
• Play / Pause media with pinky finger gesture
• Global hotkey to start/stop gesture control
• Real-time hand landmark visualization with circles and lines
• Audio feedback with beeps
• Simple gesture-based control

Tech Stack

• MediaPipe – Hand landmark detection
• OpenCV – Webcam access and image display
• pynput – Global hotkey support
• keyboard – Keyboard input simulation
• NumPy – Distance calculations
• Pycaw – System volume control (currently commented)

Requirements

Python Version

Python 3.9 - 3.11 (64-bit) recommended

Required Libraries

Install all dependencies with this single command:

pip install mediapipe opencv-python pycaw comtypes keyboard pynput numpy

Deployment / Installation

1. Clone or download this repository
2. Open Command Prompt or Terminal as Administrator
3. Install the required libraries using the command above
4. Save the main code as gesture_control.py in the project folder
5. Run the application:

python contro2l.py

Usage Instructions

After running the script, the program starts in waiting mode and prints "waiting..".

1. Press the hotkey Ctrl + Alt + Cmd (Ctrl + Alt + Windows key) to activate gesture control
2. The camera window titled "Image" will open
3. Show your right or left hand clearly to the webcam

Available Gestures

• Scroll Down: Move index finger down while thumb is close to index finger
• Scroll Up: Move middle finger up while thumb is close to index finger
• Play / Pause: Spread pinky finger far from thumb (triggers toggle with beep sound)

Exit Controls

• Press Ctrl + Alt + Cmd again to stop gesture mode (cmd is the win button in windows)
• Press Spacebar in the camera window to completely exit

Tips for Best Performance

• Ensure good lighting on your hand
• Keep your hand 30–60 cm away from the webcam
• Avoid very fast movements for more accurate detection
• If the hotkey doesn't work, try changing it in the code to <ctrl>+<alt>+h

Known Issues & Notes

• Hotkey using <cmd> (Windows key) can sometimes be unstable with pynput. Use <ctrl>+<alt>+h as backup if needed.
• Pinky detection uses landmark 18 (not the tip). This may be improved later.
• Volume control code is commented out and can be enabled easily.
• Zoom gestures are detected but not implemented yet.

Future Enhancements

• Enable volume control using thumb-index distance
• Add actual zoom functionality
• Implement mouse cursor control
• Improve gesture smoothness and add filtering