Hey Guyzz...
The research team from Stanford University have previously reported the burglary protection system obchnoy CAPTCHA, as well as audioCAPTCHA, now created an algorithm for breaking the new video technology nuCAPTCHA.
As a rule, technology CAPTCHA, or Completely Automated Public Turing Test to Tell Computers and Humans Apart is a text that a person must enter before you make any transaction on the site, for example, send your comment. This technology is designed to protect against bots or automated programs that send advertising messages on the forums, comment on blogs, etc. The problem with CAPTCHA is that if the algorithm is writing the text is too simple, bot with embedded OCR-engine ( optical character recognition) will be able to understand the coded words and phrases if the algorithm is hiding would be too complicated, it can not recognize the phrase, and people. Studies show that approximately 25% CAPTCHA-images from the first time detected and wrong person.
In NuCaptcha offers a different approach to the protection, through the use of video, which will show the letters and numbers. Such a system is not static, like a normal CAPTCHA and the bot will not recognize it. Technically, an automated pattern recognition technology video already, but they are at very early stages of development and their algorithms are far from ideal. NuCaptcha second improvement lies in the simplicity of integration of new technologies to protect the site.
For breaking new technology, researchers have created a special software that makes a lot of shots and smashes nuCAPTCHA video or animation frames. After that, the application analyzes each frame, like an ordinary picture. Also, the application uses simple background removal system to facilitate the decryption code. Today, as the developers say their system is compromised nuCAPTCHA works for 90% correct.
The creators of the program say that in some cases, the program is faltering due to moving characters.
Further details are available here on the development of: http://elie.im/blog/security/how-we-...ose-to-fix-it/
0 comments:
Post a Comment