To begin with lets explain what kind of problem code CAPTCHA is. Explanation in English can be as follows: «Completely Automated Public Turing test to tell Computers and Humans Apart». In other words it’s much easier – it helps to distinguish computers from the people. This text is sometimes very useful, moreover, it is used by a very wide audience.
What is the purpose of capture using?
For example, in all of your favorite social network Facebook it should be administered, if the number of operations for a certain period of time exceeds the allowable conditions. Roughly speaking, if you send 50 messages in a row of intervals of 1 second, then surely CAPTCHA pops up, as you will be suspected in spam. But spam is made mostly by the robots, which can not (according to the authors) enter text from the picture. Accordingly, it is a necessary measure, which protects the resources from the spam and the increased load when attacking bots.
However, captcha is not always helpful, since not for each lock the key can be found, otherwise the lock will break and become useless. Most of the drawings can be identified by neural networks, if previously coaching them on many examples (several tens or even hundreds of thousands).
On many sites to confirm that you are a real person, not a robot, it is necessary to solve the so-called “captcha” – for example, to recognize the distorted letters in the image. Among the embodiments of these systems reCAPTCHA is the most significant. The user is prompted to enter the two words which are taken from scanned books. One word is easy to read, and according to it the checking is made, and the second word is much more complex, and its accuracy is not analyzed, since it is not recognized by the automatic scanning system. These words are offered to different people, and then the system takes the option that is introduced the most frequently – thus millions of Internet users are helping computers in the digitization of old books.