/AutoLabelImg

auto-labelimg based on yolov5, with many other useful tools

Primary LanguagePythonMIT LicenseMIT

AutoLabelImg:MultiFunction AutoAnnotate Tools

AutoLabelImg

Introduction:

Based on labelImg, we add many useful annotate tools, in Annoatate-tools and Video-tools menu, including:

  • TOOL LIST
  • Auto Annotate:anto annotate images using yolov5 detector
  • Tracking Annotate:using tracking method in opencv, annotate video data
  • Magnifing Lens:helpful when annotating small objects, optional function
  • Data Agument:data agument
  • Search System:search details info based on your input
  • other tools:label selecting/rename/counting, fix annotation, video merge/extract, welcome to try

Demo:

seen in Vtuber:

Auto Annotate

Tracking Annotate

Magnifing Lens

Data Augment

Search System

Update log:

2022.01.14:remove Retinanet( matain yolov5 only), and add label selecting when autolabeling

2022.01.11:imporve magnifing lens, more fluent and can be shut

2020.12.28:add video tracking annotate

2020.12.10:autolabelimg,version 1.0

Installation:

  1. clone this repo:

    git clone https://github.com/wufan-tb/AutoLabelImg
    cd AutoLabelImg
  2. install requirments:

    conda create -n {your_env_name} python=3.7.6
    conda activate {your_env_name}
    pip install -r requirements.txt
  3. compile source code:

    Ubuntu User:

    sudo apt-get install pyqt5-dev-tools
    make qt5py3
    

    Windows User:

    pyrcc5 -o libs/resources.py resources.qrc
    
  4. prepare yolov5 weights file and move them to here: [official model zoo:Yolov5]

    mv {your_model_weight.pt} pytorch_yolov5/weights/
  5. open labelimg software

    python labelImg.py
    

Set shortcut to open software[optional]

Windows User:

create a file:labelImg.bat, open it and type these text(D disk as an example):

D:
cd D:{path to your labelImg folder}
start python labelImg.py
exit

double click labelImg.bat to open the software.

Ubuntu User:

open environment setting file:

vim ~/.bashrc

add this command:

alias labelimg='cd {path to your labelImg folder} && python labelImg.py

source it:

source ~/.bashrc

typing 'labeling' in terminal to open the software.

Citation

{   AutoLabelImg,
    author = {Wu Fan},
    year = {2020},
    url = {\url{https://https://github.com/wufan-tb/AutoLabelImg}}
}