Robotic Speed Reading
from the useful...-if-it-weren't-for-copyrights dept
As librarians are realizing the benefits of digitizing all of their works, they’re also discovering what a huge process it is to manually scan books. Character recognition software has certainly improved to the point that it’s quite useful, but the issue of manually turning pages is a huge problem. Along comes a robot to the rescue. Apparently, two different companies have designed special book scanning machines that turn the pages for you as they scan – scanning up to 1,000 pages an hour. This, of course, would be much much much more useful if we didn’t have draconian copyright laws that meant putting most books in this machine is a criminal act. Update: Oddly, the original story has disappeared from the site just a day after it came out, so here’s a slightly different version.
Comments on “Robotic Speed Reading”
link is blank for me
At least, a big blank area where the article should be.
scanning machines / Rocky Mountain News
The page comes up blank for me, too, and with a couple of minutes poking around I still couldn’t find a link elsewhere on their site.
This -may- be the newspaper’s take on the story of a few days ago about the scanner installed at Stanford, currently converting the CSLI archives into digital form. Other interesting projects include Benetech.org in Palo Alto, scanning books for the blind and limited-vision, and the venture MIT and HP put together to digitize the MIT Press backlist, using auto scanners and digital character recognition with a high claimed accuracy.
Our experience at Hidden Knowledge, in digitizing public-domain books for publication as e-books, has been that commercial OCR software is still inadequate and not getting better very fast. OTOH, as storage becomes cheaper and cheaper it becomes more reasonable to store digital images of the pages, say monochrome at 600 dpi.
Mike