Gotcha ! Google learns to read with reCAPTCHA

If you leave a comment on this page, chances are you'll face a CAPTCHA box. This "Completely Automated Public Turing test to tell Computers and Humans Apart" prevents robots from spamming. If you pass the test, you'll face a human being, Yours Truly. And since my platform is Blogger (a Google unit), CAPTCHA provider will eventually be reCATCHA, the company purchased by Google yesterday.

Each time Microsoft or Google snatches a start-up, conspiration theorists try to decypher the CAPTCHA, as in "Completely Automated Private Takeover to tear Competitors and Humans Apart".

This time, Google provides the key on its blog* : they need the technology to capture our own know how in text recognition for its own document scanning processes, most notably for Google Books, which recently claimed a spectacular partnership with one of its former opponents, the Bibliotheque Nationale de France.

Who knows ? The system could also boost other applications in the long run... Writing to speech ? At times I'd love to find a tool to decypher my own scribbles.




* "Teaching computers to read: Google acquires reCAPTCHA" (20090916) :

" Computers find it hard to recognize these words because the ink and paper have degraded over time, but by typing them in as a CAPTCHA, crowds teach computers to read the scanned text"

ADDENDUM 20090919

Since this post received yesterday a visit from Carnegie Mellon Algorithms and Complexity Theory Group, I shall add, for full recognition, that reCAPTCHA happens to be a spinoff of Carnegie Mellon University’s Computer Science Department.