Warning: Use of undefined constant fb_replace_wp_version - assumed 'fb_replace_wp_version' (this will throw an Error in a future version of PHP) in /usr/home/heximal/data/www/heximal.ru/public/blog/wp-content/plugins/replace-wp-version/fb-re_wpversion.php on line 39
The truth about reCAPTCHA | heximal blog

Warning: Cannot modify header information - headers already sent in /usr/home/heximal/data/www/heximal.ru/library/Heximal/Lang.php on line 43
Home > News > The truth about reCAPTCHA

The truth about reCAPTCHA

Recently I’ve learned the truth about reCAPTCHA – a project of Google good company. For reference, CAPTCHA means Completely Automated Public Turing test to tell Computers and Humans Apart. Alan Turing was the first person who coined the term Artificial Intelligence (AI) and developed a specific test to determine AI. So, the CAPTCHA is a special case of Turing test for artificial intelligence. Now, what I wanted to tell you about reCAPTCHA.

There are two ways to implement a web site AI protection ​​(anti-spam protection). First you can implement it yourself (which is not too hard), or second one is to use the reCAPTCHA service (which is much easier). So far I have treated with reCAPTCHA honestly with big misunderstanding: I had no idea why so many well-known web sites use reCAPTCHA, when they could have easily implement their own CAPTCHA machanisms. Besides, one can look for such a monstrosity captcha for lon long time, and indeed, why they are so proud that they have to use this public website engine – well, because the ugliness of candid… I found the truth as I said and that truth was cool.

Characters images on the picture are generated not automatically at random, as is used to be in classical captcha. Google’s has a global OCR project (OCR – Optical Character Recogintion) – thousands and thousands of printed text pages, scanned and sent to the automatic processing (translation of the bitmap image to the text representation). OCR task belongs to a barnch of artificial intelligence, and there is still no sane recognition algorithm, which would give a high success rate. There are always unrecognized fragments remain, which must be processed manually. Google has a brilliant solution to join two projects. All that is not recognized by scanning robo is passing through a network of sites, using the second service (reCAPTCHA) to users of most Web sites who perform this work manually. Thus, the whole machine is fully automated. Among the rest of bonuses Google gets a free teacher for a neural network of his artificial intelligence that deals with recognition.

And that’s not about it. reCAPTCHA has another side – speech recognition. If you look at the reCAPTCHA frame, in addition to the text where you can find a button with a speaker icon. The idea is exactly the same, but deals with sound recognition.

Categories: News Tags:
  1. August 12th, 2012 at 01:24 | #1

    Если не ошибаюсь, одно слово сгенерировано системой, а второе взято из текстов, отправленных для распознавания.

  2. August 12th, 2012 at 02:02 | #2

    В точности так.

  3. March 7th, 2013 at 02:15 | #3

    действительно круто придумано