https://github.com/jshulgach/hand-landmark-tracker

An app created with PyQt5 using Google's MediaPipe hand landmark ML model for hand tracking and streaming
https://github.com/jshulgach/hand-landmark-tracker

computer-vision hand-tracking machine-learning mediapipe opencv pyqt5 python

Last synced: 9 months ago
JSON representation

An app created with PyQt5 using Google's MediaPipe hand landmark ML model for hand tracking and streaming

• Host: GitHub
• URL: https://github.com/jshulgach/hand-landmark-tracker
• Owner: Jshulgach
• Created: 2023-08-13T15:09:16.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2025-06-14T00:18:12.000Z (about 1 year ago)
• Last Synced: 2025-10-09T15:33:25.044Z (9 months ago)
• Topics: computer-vision, hand-tracking, machine-learning, mediapipe, opencv, pyqt5, python
• Language: Python
• Homepage:
• Size: 72.4 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          


  



# Hand Landmark Tracker  [![](https://img.shields.io/badge/python-3.8.5-blue.svg)](https://www.python.org/downloads/)

Hand Landmark Tracker bring a new perspective to human-computer interaction using hands. The hand is the most versatile and intuitive controller someone can use, so it makes sense to see if there is a way to design an interface that takes advantage of the hands without requiring them to touch anything.

This code uses the amazing features of Google's machine learning suite [MediaPipe](https://developers.google.com/mediapipe), a media-based ML package for classification and recognition with neural networks.

## Installation

1. Create a virtual environment using [Anaconda](https://www.anaconda.com/products/distribution) or Python's virtualenv.

   - Using Anaconda:

      ~~~

      conda create -n hand-tracker python=3.8

      conda activate hand-tracker

      ~~~

   - Using Python's virtualenv:

     ~~~

     python3 -m venv hand-tracker

     source hand-tracker/bin/activate # Linux

     call hand-tracker/Scripts/activate # Windows

     ~~~

2. Clone the repository and navigate to the project directory.

   ~~~

   git clone https://github.com/Jshulgach/Hand-Landmark-Tracker.git

   cd Hand-Landmark-Tracker

   ~~~

3. Install the required packages.

    ~~~

    pip install -r requirements.txt

    ~~~

## Getting Started

### Hand tracker Class

The HandTracker class can be easily imported and used in your own projects by creating an instance of the class and calling the `track` method.

```python

from tracking import HandTracker

tracker = HandTracker(visualize=True) # Enable visualization on a CV frame

tracker.run()

```

This will start hand tracking and displaying any landmarks captured on top of detected hands.



  



### MiniArm Gripper Virtual Spacemouse Control

A more interesting demo is using hand tracking like a virtual spacemouse as a position displacement controller operating the position of a robot end effector. We can use the [MiniArm](https://github.com/Jshulgach/Mini-Arm) robot for example and send position and gripper commands to the robot server.

To run the spacemouse demo, the `main.py` script can be run with the `spacemouse` mode and COM port set. Enabling the verbose mode allows for text output of the tracking data.

~~~

python main.py --mode spacemouse --port COM5 --verbose true

~~~



  





  



## Acknowledgement

The project was inspired by xenon-19's [Gesture Controlled Virtual Mouse](https://github.com/xenon-19/Gesture-Controlled-Virtual-Mouse) project.

 Thanks Google for the amazing [MediaPipe](https://developers.google.com/mediapipe) library.