Captcha at Work: Re - Captcha Interpreting Scanned Documents

[Chuckiechan]

What is reCAPTCHA?

Many old documents are being scanned into the public domain, and due to hand set type, stains, mold, blood , and other "issues", the scanners can't come out the accurate text conversions.

So when you enter "captcha" you see a second word ("re captcha"). What you actually could be doing is correcting a scan of a 1000 year old document!

The theory is that if you can do the "captcha" accurately, you will do the "re-captcha" accurately, also. When the collection of results leans far enough to one word, it is considered to be correct and in inserted into the document.

Very clever...

[surreal]

I saw that guy interviewed on the PBS science show. They are just burning though stuff getting it digitized now.

[Chuckiechan]

I think that is what I saw the made me want to dig it up to share.

Was he some super bright guy now a professor?

[surreal]

Yes, and still a kid (like 24 or something)

[tony_j15]

30, actually. Math nerds age well

His name is Luis von Ahn.

Nova Science Now was the show, here's the segment.

[surreal]

Quote:

Originally Posted by tony_j15

30, actually. Math nerds age well

It's amazing how nice the skin stays without any exposure to sunlight!