Licensing opportunity
Researcher Hung-Son Le at Umeå University, in Sweden, has developed algorithms that can carry out facial recognition on the basis of one photograph only.
Currently facial recognition systems rely on a large collection of images of each face, captured at different levels of illumination and in different poses. Acquiring the images for each individual is difficult and expensive. Moreover the sensitivity of the system can be compromised by poor quality photographs, different facial expressions, angles and levels of illuminations.
The algorithms developed by Hung-Son Le make it possible to identify a face even if there is only one picture in the database for each person. The technique involves improving the contrast in underexposed and overexposed pictures, so that details are made visible which it would otherwise be difficult for the computer to identify.
Tests against international standard systems have demonstrated that it outperforms its leading competitors.
Commercial applications are under development, including, a face web search engine to demonstrate the technology. This will soon be accessible via links from Hung-Son Le’s home page at http://www.tfe.umu.se/personliga/slh.
Receive the Funding Newswire each Tuesday, our Policy Bulletin each Thursday, and news about bridging Europe’s east-west innovation gap twice a month in The Widening. 