Digital Archivists 2014-05-29

Attendees[edit]

John gave an update on contacts with the DPLA service centers and other possible partners
Vinay shows portable Freedom Scientific Pearl scanner for the blind and begins discussion of problems with its design and price point
Vinay summarizes two OCR programs: Finereader has better OCR and is preferred by libraries, Omnipage is generally better at understanding layouts
Pearl scanner price points: $800 for hardware, $800 for software
Commercial OCR programs good at batch processing
Companies that make accessibility products price-gouge because they serve a small niche and are usually paid for by government funding
Discussion of Pearl scanner design flaws: Power cord bound to short-circuit because of case design, non-functional LED, impossible to adjust height, camera is only 5MP, software is way more expensive than it needs to be (OCR licenses themselves are only $40 or so)

* RESOLVED: We will gather hardware components, possibly 3D-printed, to build a better portable scanner

John enlists friends to bug Miloh to see if he can help with this
Andrew and Vinay compare OCR platforms
Dany shows how to use diff and meld to compare OCR results
Discussed document scanning with iPhones
Vinay discusses flaws in software that comes with Pearl scanner, offers to code a better version for our new portable scanner
Discussion of frames for iPhone scanning
General discussion of improving OCR through algorithms: NLR, language dictionaries, font differences, democratic algorithms, dominant languages
Dany and Andrew dig into Tessaract configuration
Andrew and Brandon set up the scanner, do some maintenance, and begin scanning books
Andrew notes that compasses should be etched into the camera and book mount
We show a few people who are new to Noisebridge the scanner
Discussion with the 3D printing group -- Ron and Dave (??) -- who are working next to us
Discussion of the Internet Archive $10K book scanner
Discussion of current book scanning services and how we can better serve libraries and others who don't want to rip out the spines but also don't want to pay 10 cents per page
Everyone goes home except Andrew and Brandon, who scan late into the night