An Open Scanner
A Creative Commons Venezuela proposal for a low-cost, open book scanner that starts from the software, so that anyone with a camera can digitise a library.
Why start from the software
In 2020, in the middle of the pandemic, the Creative Commons Community Activities Fund accepted a proposal from the Venezuelan chapter: build a version of an open-source book scanner. The first audience was obvious to us, university libraries full of volumes that exist only on paper. The plan was to replicate an existing open design and start cutting and pasting, as others had done.
It did not work that way. The open hardware designs were excellent; the software that made them useful was hard to find, use or maintain, and each scanner was built around one way of taking and processing the pictures. So we started from the cheaper half, the software, with a different question: if we do not care how the photographs are taken, can we still reconstruct the document and read it?
Exhibit 1DiagramRead left to right
A scanner that starts from the software
The hardware is whatever camera and frame you have. The intelligence is in the software, and the people are split into two roles that do not need to share a room: Operators who photograph, Editors who reconstruct.
Operators and Editors
Computer vision and machine learning answered yes: pages photographed with a commodity camera can be reconstructed and passed through optical character recognition, offline or as a service in the cloud. That one decision decouples the hardware from the software. Any camera will do, any frame can be built or borrowed, and the cost of the scanner stops being the barrier.
It also creates two kinds of people. Operators carry the scanner, take the photographs and train the next users; Editors, who need not be in the same city or country, reconstruct the books and run the recognition. Scanning in one place and editing in another is how a small chapter could digitise many libraries, and how the work could become paid work where the internet is scarce. In 2020 we built version zero of the hardware and the software; Estefanía Sánchez Pineda gave the project its identity under a CC BY licence; and in March 2021 I wrote the first part of the story for the Creative Commons network.
Exhibit 2TimelineFilled points are my milestones; hollow points are the context they sit in
From a grant to a write-up
- 2020The Community Activities Fund accepts the proposalA task force forms in Venezuela and abroad; university libraries are the first target.Source
- 2020Version zero of hardware and softwareA commodity camera, a simple frame, and the reconstruction working from photographs.Source
- 2021Part 1 published, 16 MarchIn English and Spanish, in the Creative Commons network publication.Source
Exhibit 3PeopleCredit where the work was shared
Credits
- Estefanía Sánchez PinedaIdentity and logo of An Open Scanner, released under CC BY 4.0; Creative Commons Global Network member.
Exhibit 4ResourcesOpen the thing itself, not a description of it
Where to go from here
- ArticleAn Open Scanner, part 1, in Creative Commons: We Like to Share, March 2021medium.comAuthor
- ArticleAn Open Scanner: una propuesta del capítulo Creative Commons en Venezuela, the Spanish versionartfisica.medium.comAuthor
- RecordThe Creative Commons Global Network Community Activities Fundnetwork.creativecommons.org
- SiteCreative Commons Venezuelave.creativecommons.netCo-founder
Exhibit 5Looking backOne dated note from inside the story
Looking back
2021, when it happened
Part 1 ended with a promise of a Part 2: clean repositories, a website for the blueprints, the stories of the first Operators and Editors.