Skip to content

Latest commit

 

History

History
76 lines (63 loc) · 3.69 KB

File metadata and controls

76 lines (63 loc) · 3.69 KB

Unity_HandTracking_DeepLearning

About project

Realtime HandTrack

1 2

Game Playing

1 2 3 4 5 6
1번 2번 3번 4번 5번 6번
  • This project is a Unity3D Project
  • You can draw pictures using certain hand gestures.
  • A trained deep learning model will tell you what your drawing is.

Unity

How to run (the built game is in the unity_build folder)

  1. First, ensure you have Git LFS.
    git lfs install  
    
  2. Then, clone this repo.
    git clone https://github.com/jiho-00/Unity_HandTracking_DeepLearning   
    
  3. All of project files are in unity/Assets. This folder includes all scripts and assets to run the project excluding python files.

Unity demo scenes

Scene name Scene Scene
Main image image
Room image image
Drawing image image

Python

Dataset

How to run

  1. Preparing your datasets

    ├── data
      ├── txtfiles
      │   ├── 30categories.txt
  2. Making train/test datasets

    python DataUtils.download_data.py
    
  3. Main Training

    python main.py
    
  4. Inference & Socket communication

    python InferenceOneImage.py