Machine learning attacks against the asirra CAPTCHA

The Asirra CAPTCHA relies on the problem of distinguishing images of cats and dogs (a task that humans are very good at). The security of Asirra is based on the presumed difficulty of classifying these images automatically. In this talk, we describe a classifier which is 82.7% accurate in telling apart the images of cats and dogs used in Asirra... We also investigate the impact of our attacks on the partial credit and token bucket algorithms proposed... One contribution of our work is to inform the choice of safeguard parameters in Asirra deployments.