Home

Welcome to the unshred-tag wiki!

Unshred Tag is Flask/Mongo application which complements unshred project.

Unshred is a library that allows to analyze scanned sheets with shreds (just like those)

Unshred Tag is a workplace for volunteers to crowdsource further processing of shreds. As some information in shreds is too complicated to be extracted automatically we decided to create our own Mechanical Turk just for this purposes.

Unshred tag can process those scanned page with help of Unshred library and load all the shreds (and their features) from scanned sheets of paper to MongoDB.

Then users can log in to the system using their favorite social network and manually tag each shred with tags like "has white background", "has numbers", "has blue ink marks" etc. We are also going to provide a web interface so user can not only tag but determine rotation and also attach recognized text to each shred.

All processing is done with controlled level of redundancy so each shred will be processed by at least N users.

Unshred tag is built on top of Flask, Mongo and Twitter Bootstrap. It's also using jQuery for frontend goodness.

Instance of Unshred tag can be deployed on Heroku (at least for now).

Unshred tag has russian web-interface, but I'm happy to help you to translate it to English if you are interested in such job.

If you want to help us with development — here is our docs on collaboration and used terminology.