Using discrete wavelet transform and eigenfaces for recognizing avatars faces

Abdallah A. Mohamed, University of Louisville
Roman V. Yampolskiy, University of Louisville


It has been observed that the security of virtual worlds is becoming an important issue for law enforcement agencies. Accurate and automatic tracking of people in the virtual world is an interesting problem for terrorism and security experts. One way to track and recognize the identity of a person in the virtual world is to recognize the face of the avatar that represents this person. To deal with this problem two algorithms capable of recognizing avatar faces with a high degree of accuracy are described in this paper. Experiments conducted on two datasets of avatars from the Second Life and Entropia Universe virtual worlds show the effectiveness of using these algorithms for recognizing avatar faces. © 2012 IEEE.