Skip to content

Latest commit

 

History

History
92 lines (66 loc) · 3.75 KB

README_EN.md

File metadata and controls

92 lines (66 loc) · 3.75 KB

HuaHuoLabel

HuaHuoLabel is a pure Python project. It is developed with PySide6 and compiled with Nuitka. It is a multifunctional AI data label tool, which supports data label of five computer vision tasks, including single-category classification, multi-category classification, semantic segmentation, object detection and instance segmentation. HuaHuoLabel can also do image editing, dataset management, auto-labeling, and pseudo label generation, which can help train AI models more conveniently.

Support OS

Windows10 (tested), Windows11 (untested)
Ubuntu20.04 (tested), Ubuntu22.04 (tested)

Support Language

中文简体, English

Highlights

  • Good user interface and operation method
Single Category Classification Instance Segmentation
cls_example.gif ins_example.gif
  • Ring form and pixel-level label
Ring form label Pixel-level label
ring.gif pixel.gif
  • Real-time image enhancement

  • Label statistics and management

  • Divide train set and validation set

  • Auto-labeling
    (to be done)

  • Pseudo label generation
    (to be done)

Usage

HuaHuoLabel usage introduction

Run Project

  1. Download the release
    Unzip the file, for Windows version, directly run the HHL.exe.
    For Ubuntu version:

    cd HHL.dist
    ./HHL
    
  2. Clone the source code and then compile it
    This project is compiled with Nuitka. Nuitka supports Windows、Linux and macOS. Theoretically, the source code can be compiled to the execution program for all the adaptive systems.

    # install nuitka
    pip install nuitka
    # For Windows
    python -m nuitka --mingw64 --standalone --plugin-enable=pyside6 --output-dir=out --windows-disable-console --windows-icon-from-ico=images/icon.png HHL.py
    # For Ubuntu
    # Because of (https://nuitka.net/info/debian-dist-packages.html), please use pip to manage python packages.
    # install patchelf
    sudo apt install patchelf
    python -m nuitka --standalone --plugin-enable=pyside6 --output-dir=out --windows-disable-console --windows-icon-from-ico=images/icon.png HHL.py

    After compilation is done, copy "images" folder, "ts_files" folder, "ui_files" folder and "project.json" to the project root directory.

  3. Clone the source code and run HHL.py in your IDE.

Something Need Help

  1. If you are familiar with PySide6, these problems need help.
    https://forum.qt.io/topic/141592/can-not-move-horizontalscrollbar-to-the-rightmost-side
    https://forum.qt.io/topic/141742/how-to-translate-text-with-quiloader

  2. Due to the author's limited English, any suggestion of the translation of the English version release and USAGE_EN.md is welcome.

License

HuaHuoLabel uses GNU GPL license. You can use this tool at will. However, do not use it for commercial activities without the permission of the author.