/labelify

labelify is a web tool for annotating and labeling object bounding boxes in images

Primary LanguageHTMLMIT LicenseMIT

Labelify

LabelImg is a graphical image annotation tool. Just like labelImg but web and responsive to easily annotations with smartphones

logo

Requirements

  • Node.js > 14.*
  • Mysql database. You could use docker
  • Create a database and execute this ddl to create the required tables: ./database/ddl.sql

Settings

key sample value description
PORT 8080 change the port of server. Default is 2708
LABELIFY_DATABASE_HOST 10.20.30.40 some ip of your mysql database
LABELIFY_DATABASE_USER usr_labelify the mysql user for your my database. Root is not required.
LABELIFY_DATABASE_PASSWORD ***** password related to the mysql user
LABELIFY_DATABASE_PORT 3306 mysql port. Default is 3306
LABELIFY_DATABASE_NAME labelify name of your mysql database
TOKEN_SECRET ***** secret for jwt generation

For developers

npm install
npm run dev

For servers

npm install
npm run start

For servers with docker

Complete details here

Usage

Open your browser pointing at:

If no errors, you will see the Login

An after login an image will be ready for your human box annotation:

home

Just perform the annotation just like labelImg (mouse drag)

Press submit, the box coordinates will be sent to the database

And then a new image will be loaded

Users

admin with password 123456 and demo uuid 9fd1d411-35db-4c08-88c3-4001d18d3cb3

Roadmap

  • unit test, coverage, badges
  • password hash
  • admin first password feature
  • user crud
  • images crud
  • annotation view
  • fba protection

Made with

Contributors

Thanks goes to these wonderful people :


JRichardsz