Face reading software

26 February 2008

Researchers in Spain have developed an algorithm capable of reading facial expressions from video images.

 

By applying the algorithm, the system is capable of processing 30 images per second to recognise a person’s facial expressions in real time before categorising them as expressing: anger, disgust, fear, happiness, sadness, or surprise.

By dividing a person’s face into ‘boxes’, facial movements of different parts of the face are analysed and compared with the expressions captured in a database. Using this database the system has a success rate of 89 per cent and is capable of working in poor conditions where ambient lighting, frontal facial movements or camera displacements cause major changes in facial appearance.

The software was created by a collaboration between the Department of Artificial Intelligence at the Universidad Politécnica de Madrid’sSchool of Computing and the Universidad Rey Juan Carlos, also in Madrid. Despite being at the prototype stage, the software is able to work on regular computers and laptops.

According to the team the technology has a range of applications including advanced human-computer interfaces where it could be used to construct avatars that simulate a person’s facial expression.