Learning Visual Features for the Avatar Captcha Recognition Challenge

Abstract

Abstract-Captchas are frequently used on the modern world wide web to differentiate human users from automated bots by giving tests that are easy for humans to answer but difficult or impossible for algorithms. As artificial intelligence algorithms have improved, new types of Captchas have had to be developed. Recent work has proposed a new system called Avatar Captcha, in which a user is asked to distinguish between facial images of real humans and those of avatars generated by computer graphics. This novel system has been proposed on the assumption that this Captcha is very difficult for computers to break. In this paper we test a variety of modern visual features and learning algorithms on this avatar recognition task. We find that relatively simple techniques can perform very well on this task, and in some cases can even surpass human performance.

A new CAPTCHA based on facial expression recognition

“…More importantly, these CAPTCHAs still have AI problems regarding image perception and interpretation 9 . Korayem et al approached the challenge from another perspective 10 , asking users to identify avatars from a set of real faces. His idea is that humans can easily recognize cropped or low-quality images of faces, which can be challenging for machines.…”

Solving Avatar Captchas Automatically

“…This paper is an expanded version of a preliminary paper[13] that appeared at the ICMLA Face Recognition Challenge[19].…”

Avatar captcha : telling computers and humans apart via face classification and mouse dynamics.

“…Users are unable to access the image databases. However, through a challenge presented at the ICMLA 2012 conference [112] to classify human and avatar facial images, it was learned that this problem was solved successfully [113,114,115,116]. We aim to generate human and avatar datasets dynamically, obtained in real-time, from popular online websites such as Flickr and ActiveWorlds.…”