Gesture-based CAPTCHA

This project aims to develop a new CAPTCHA scheme (Completely Automated Public Turing test to tell Computers and Human Apart) with enhanced mobile user experience while retaining the necessary level of security. Unlike conventional text-recognition based CAPTCHA schemes and even the latest interactive CAPTCHA schemes, this proposed CAPTCHA scheme takes a hybrid approach to strike a balance between usability and security.

The scheme features an instructional layer which utilises cognitive psychology findings to form a CAPTCHA challenge which can be easily understood by human users but not by bots and an interaction layer which allows users to perform gesture friendly interactions to respond to the CAPTCHA challenge. A typical example is to partially process object moving challenge instructions to force a user to utilise context effect to understand the instruction and respond to it. Since computer bots cannot easily apply context effect to understand the instruction, they have to use brute force to try out all possible movements. Then the security can be retained by simply adding more success criteria to lower the probability of using brute force to crack the challenge.


Contacts: Nan Jiang

Funders: Bournemouth University


Related Publications