Facial-recognition software has been created for mobile phones that could soon help identify individuals and how they are feeling.
The system has been developed by researchers at Manchester University based on existing facial-recognition technology used on computers. It tracks 22 facial features in real time and is thought to be unrivalled for speed and accuracy.
‘Trackers used in mobile phones today do not provide a clear understanding of the position of the face,’ said Dr Phil Tresadern, lead researcher on the project. ‘Our tracker uses a combination of voice and facial-recognition software making it far more effective.’
The software uses the camera on a mobile to track individual points on a person’s face, focusing on the eyes, noise, mouth and jaw line. The image is stretched to provide a neutral expression, which is then compared to a database of existing photos. A user’s voice is also recorded to help identification.
‘The software uses a very small percentage of the memory of a mobile phone,’ said Tresadern. ‘Advances in hardware and the display technology have meant that you can get a good picture back, which had previously not been possible.’
Tresadern believes facial-recognition technology will soon replace passwords and PIN numbers to log into internet sites from a mobile. But he also claims it can be used for novel ways of interacting with a mobile phone, such as for augmented reality applications or to record a user’s emotions.
The software has been demonstrated on a Nokia N900 for the EU-funded Mobile Biometrics (MoBio) project and has attracted interest from mobile phone manufacturers. It is currently undergoing further development to improve its use under different lighting conditions.