Share
Blog
Cite
Print
Email
BookmarkScienceDaily (Jan. 28, 2008) — Hung-Son Le, Umeå University, Sweden, has developed in his PhD dissertation the algorithms that give a computer the possibility of recognising a face, even if only one picture is taken. The results can be used for safe and secure identity control or, on the light side, to find out to which famous persons you look alike.
See also:
If a non-authorised person gets access to your pin code and credit card, most likely your money will disappear from your account. Nevertheless, this would be impossible if the ATM could recognize your face as you look at a camera. Now, the algorithms to carry out this function, face recognition, exist. Face recognition can also be used in other functions, for instance in a dating service. Maybe the customer is interested in a man that looks like Brad Pitt or a woman that looks like Angelina Jolie.
Systems that can identify different faces are normally trained through a database with a large collection of face images in different illumination and pose. Nevertheless to collect such a large number of face images for each person is difficult and quite often expensive. Moreover these systems have problems due to the bad quality of the pictures, as well as facial expressions, the variety of angles and the different illuminations. These problems are now over.
The effective algorithms developed by Hung-Son Le make it possible to have a system that can identify a face even when there is only one picture in the database for each person. Moreover, the effectiveness of the system is a considerable improvement when taking into account light conditions, or facial expressions. His algorithms use a method than improves contrast in underexposed and overexposed pictures. Thus details can be made visible which otherwise would be difficult for a computer to identify. Given the method used (Hidden Markov Model, HMM), once the system is in place, it needs no time for retraining, when compared to existing HMM-based competitors, to “know” new pictures with different expressions taken under different illumination conditions.
The experiments carried out with the system and tested against international standards such as FERET and the Yale database, have demonstrated that it outperforms the leading competitors.
Commercial applications based on the PhD dissertation results are under development and will soon be presented. Among others, a face websearch engine is under final development phase.
