Skip to content

Latest commit

 

History

History
43 lines (27 loc) · 1.46 KB

README.md

File metadata and controls

43 lines (27 loc) · 1.46 KB
Raw

3D Hand Tracking Using Mediapipe and Unity

In this repo I used Mediapipe package for extracting landmark from a hand and then I send these landmarks data to the Unity using socket and then I show the landmarks and their connections.

How to install

Run this command

pip install -r requirements.txt

Then you need to install Unity. You can download and install it from here.

How to run

First you need to copy my Assets folder into your Unity project instead of your Assets folder

Then you need to run your Unity project

Then you can run the AI using the following command

clear; python main.py

You can also see the other arguments of it with this command

clear; python main.py --help

For Example:

  • --webcam: You can change your webcam ID. default:0
  • --min-conf: You can change the minimum number of confidence for hand detection. default:.7
  • --min-track-conf: You can change the minimum number of confidence for hand tracking. default:.8
  • --ip: You can change the IP address to send data to Unity and C#. default:127.0.0.1
  • --port: You can change the port to send data to Unity and C#. default:5052

Result

Result