labelmeWithAI

labelme: Image Annotation Tool with Python

Labelme is a graphical image annotation tool inspired by http://labelme.csail.mit.edu.

It is written in Python and uses Qt for its graphical interface.

Dependencies

• PyQt4

Installation

Docker

You need install docker, then just run below:

wget https://raw.githubusercontent.com/wkentaro/labelme/master/scripts/labelme_on_docker
chmod u+x labelme_on_docker

# Maybe you need http://sourabhbajaj.com/blog/2017/02/07/gui-applications-docker-mac/ on macOS
./labelme_on_docker _static/apc2016_obj3.jpg -O _static/apc2016_obj3.json

Ubuntu

sudo apt-get install python-qt4 pyqt4-dev-tools
sudo pip install labelme

OS X

brew install qt qt4
pip install labelme

macOS Sierra

# on python2
brew install pyqt5 --with-python
pip install git+https://github.com/wkentaro/labelme.git@pyqt5

# on python3
brew install pyqt5
pip3 install git+https://github.com/wkentaro/labelme.git@pyqt5

Usage

Annotation

Run labelme --help for detail.

labelme  # Open GUI
labelme _static/apc2016_obj3.jpg  # Specify file
labelme _static/apc2016_obj3.jpg -O _static/apc2016_obj3.json  # Close window after the save

The annotations are saved as a JSON file. The file includes the image itself.

Visualization

To view the json file quickly, you can use utility script:

labelme_draw_json _static/apc2016_obj3.json

Convert to Dataset

To convert the json to set of image and label, you can run following:

labelme_json_to_dataset _static/apc2016_obj3.json

Sample

• Original Image
• Screenshot
• Generated Json File
• Visualized Json File

Screencast