AVATAR AUTHENTICATION

We first recall techniques for collecting and classifying databases of avatars and bots, described in Gavrilova and Yampolskiy (2012), moving on to propose a new way to synthesis the new images through application of biometric synthesis methods based on geometric processing and multi-resolution techniques. We then study the two main types of authentication in virtual world: visual and behavioral, and introduce the multi-modal system for enhanced performance.

As was pointed out before in this book, there are mainly three different types of biometric databases: true database, virtual database and synthetic database. For unimodal biometrics, there is an abundance of freely available datasets with vendor competitions often being held and benchmarks on recognition algorithm performance being established. This is not the case in the area of virtual reality. Labeled public datasets of avatar faces, robot faces, or attributed conversations from artificially intelligent agents are currently unavailable. Some recent papers attempted to tackle the problem by looking at applying methods for face generation gender and human versus but classification to a virtual reality domain.

Cyber security, without a doubt, is one of the key concerns of many modern organizations as well as private citizens worldwide.

There have been a number of attempts by malicious intelligent software to obtain unlawful access to information or system resources. It affects security of virtual communities, social networks, and government supported cyber-infrastructures. Employing new methods to counter those threats is one of the goals of biometric and cyber security.

Extensive research on behavior based profiling of software agents as an unobtrusive way of separating helpful bots from. Additional research in artimetrics is likely to produce more behavior-profiling methods specifically designed to take advantage of the unique structure of artificially intelligent programs.

This review chapter describes a new subfield of security research which transforms and expands the domain of biometrics beyond biological entities to include virtual reality entities, such as avatars, which are rapidly becoming a part of society. Artimetrics research at Cyber security Lab, University of Louisville, USA and Biometric Technologies Lab, University of Calgary, Canada, builds on and expands such diverse fields of science as forensics, robotics, virtual worlds, computer graphics, biometrics and security. The chapter discusses how verification and recognition of avatars can be ensured by analyzing their visual properties and behavioral profiling. It also introduces a multimodal system for artificial entities recognition, simultaneously profiling multiple independent physical and behavioral characteristic of en entity, and creating a new generation multimodal system capable of authenticating both biological (human being) and non-biological (avatars) entities.