Image annotation for the machine learning community

[mattpiz]

• Application: https://annotation-app.pizenberg.fr/
• Documentation: https://reva-n7.gitbook.io/annotation-app/
• Source code: https://github.com/mpizenberg/annotation-app

I’ve been asking many questions on slack the past few weeks so I wanted to share the result of this work, and thank all the persons that were super helpful!

This application is (fully) 95% Elm, 5% JS and Co. It is intended to be useful to the machine learning community in image processing where annotated datasets are required to train models. I hope it can also serve as an example of an elm application.

The technical part of the documentation is not ready yet. But if you are interested, it is based on style-elements for the layout, and using various packages of the community including notably:

• mdgriffith/style-elements for the layout
• opensolid/geometry -> now under ianmackenzie/elm-geometry for the base geometry constructs
• zwilias/elm-rosetree for manipulation of the classes in the left sidebar
• mpizenberg/elm-pointer-events for supporting pointer events

Any feedback welcomed, and if it is in your center of interests, I’ve labelled few issues with “good first issue” for potential contributors.

[mattpiz]

I finished writing the technical part of the documentation. It’s actually more a code architecture overview than a deep dive into all the code, which makes more sense IMO. It includes a brief introduction to TEA with diagrams like this one:

There is also a short contribution section for potential contributors.
Cheers

EDIT updated diagram:

[rtfeldman]

Very cool!

I’m not sure if it would help clarity, but it would be accurate to draw a line from Model down to Sub with subscriptions on the side of the line (like how there’s a line from Model to View with view on the side of the line).

[mattpiz]

Good catch! I so rarely use the model in my subscriptions that I forgot about it