CHGR: Captcha generation using Hand Gesture Recognition

Abstract

In today’s world, the advanced technology is deployed in almost every field, it is therefore important to put efforts in security as well. CAPTCHAs were invented to ensure the security by confirming that the user using a machine is human or not. CAPTCHA is simply a test that tells human and the computer program apart. It is one of the important components that almost every commercial website include to determine whether the one using website is actually a human or not. Till date, most used CAPTCHAs are usually alphabets or numbers or combination of two. CAPTCHA can be solved by human beings only and not by bots and thus helps in stopping hackers and especially their crawlers also called spiders from acceding un-perplexed areas of the net. CAPTCHAs used till date have now become time-consuming and need more human efforts so that program could not crack it and this has become a major issue to consider as lead to time wastage. This paper has proposed a better model to generate CAPTCHA with the help of hand gesture recognition techniques which is more efficient than the traditional CAPTCHAs. In this method, a form has been displayed containing one CAPTCHA image showing some hand gesture from the database with a message displayed for ‘the user to copy the same gesture, the user then with his/her conscience has made some gesture in front of the camera of the system and then the model has verified the gesture made by the user is same as that shown or not. Following this method will surely ensure the identity of the user as it is not possible for the robot to understand by own how to copy the same gesture as shown in the image. These CAPTCHAs can be used in future by the websites to check whether the one using the website is a human being or a robot fulfilling the requirement of less human efforts and time. The experiments give satisfactory results of the proposed model.