https://github.com/smileinnovation/imannotate

Image annotation tool to make Machine Learning or others stuffs
https://github.com/smileinnovation/imannotate

image-classification tool

Last synced: 28 days ago
JSON representation

Image annotation tool to make Machine Learning or others stuffs

• Host: GitHub
• URL: https://github.com/smileinnovation/imannotate
• Owner: smileinnovation
• License: apache-2.0
• Created: 2018-06-12T08:36:26.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2023-01-07T02:24:53.000Z (almost 2 years ago)
• Last Synced: 2024-08-04T03:13:19.121Z (4 months ago)
• Topics: image-classification, tool
• Language: TypeScript
• Homepage: https://smileinnovation.github.io/imannotate/
• Size: 3.53 MB
• Stars: 45
• Watchers: 7
• Forks: 12
• Open Issues: 30
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-data-annotation - imannotate - - image (Image / video / Open source)

README

        
![](/ui/src/assets/Logo.png)

# Image annotation for Machine Learning

Convolutional Neural Networks require annotated images to learn how to get recognize objects in images, which means that you need to give both the image and _bounding boxes_ of each object to recognize when you train your model. And you will need **a lot** of annotations to ensure a good train!

Imannotate is a tool to help you to create those kind of dataset by spreading a dataset of non annotated pictures across several users and let them annonate them. Once done, you can export a CSV with all your annotations to train your model.

## Features 

- Manage multiple project

- Assign multiple users to one or multiple projects

- Annonate (indeed) by drawing bounding boxes and choose a tag

- Support "nothing to tag" on a picture

- Support "Pass" a picture

- S3 connector for the pictures source

- Qwant demo connector included (not suitable for production)

## Installation

Since we are including a `docker-compose.yaml` file, the quickest way to install and start the application is to do a `make prod` into your terminal, in the project root folder. 

However,  for the one who like to do it by themself, we have a full _Build your own_ documentation [available into the project's wiki][installation]. 

## Contributing

There's just not one way to contribute to a project and you don't have to be a developer to contribute. If you are a developer, you may want to check our documentation regarding [setting-up your environment to be able to code][contributing-dev]. Otherwise you can:

- [Filling issues][issues]

- Automating Workflows

- Writing documentation & runbooks

- Thinking about long term priorities, [doing feature request][issues]

- ... 

![](https://cl.ly/f66c139391df/non-coding-contributions-signed.png)

## Licence

The license [can be seen here][license]. Please note that we'll enforce signed-off commits and [DCO](https://github.com/integration/dco). 

![](https://cl.ly/d1720d2e0243/sponsoredby_smileinovation%252520-%252520small.png)

Opensourced by [Smile](http://smile.eu) with ❤️, made by all our wonderfull contributors.

[contributing-dev]: https://github.com/smileinnovation/imannotate/wiki/Contributing

[installation]: https://github.com/smileinnovation/imannotate/wiki/Installation

[issues]: https://github.com/smileinnovation/imannotate/issues

[license]: /LICENSE