News Archive

A dot-sized part of a face may soon be all that is needed to identify a person, according to a face recognition expert at The University of Western Australia.

Ground-breaking research by Associate Professor Ajmal Mian is investigating how to use satellite technology to identify facial features that lie under the skin.

"Multi-spectral imaging can be used to measure light reflected off a face at hundreds of discrete wavelengths in the visible spectrum and beyond," he said.

Professor Mian, an Australian Research Fellow who has worked in the field for eight years, said his research may also be able to detect people who have used cosmetic surgery to alter their looks.

"Recognition based on sets of facial images from surveillance cameras, YouTube videos, Google Images or personal photo albums is more accurate because they contain more information," he said.

"Face recognition technology is being used increasingly for computer log-ons, identity checks and surveillance, and is a boom industry around the world.

"It can be used in any kind of machine such as mobile phones, computers and robots.  It's the most user-friendly way to authenticate someone and is now so sophisticated that machines can identify a face no matter what the expression.

"Humans are very good at finding a familiar face in a crowd but less able to identify someone they may have seen only once.  This is where machines outperform people because they can memorise images and never tire of matching them to faces in a crowd."

Associate Professor Ajmal said face recognition technology was better than fingerprinting because it didn't require special equipment or an expert to verify the results.

Also, any part of a face could be used, and many images of a person's face - including different expressions and poses - could be merged to make a composite image which was more meaningful to the machine.

"Humans can recognise a person regardless of whether they're laughing, frowning, crying or sleeping.  Machines may soon be able to do the same."

Associate Professor Mian is the only West Australian to have won the Australasian Distinguished Dissertation Award from CORE (The Computing Research and Education Association of Australasia).

He has won two prestigious national fellowships: the Australian Postdoctoral Fellowship and the Australian Research Fellowship, and has written more than 50 high-impact papers including more than 30 as first author.

He has 563 Google Scholar citations, including 148 on his most highly cited paper.

Media references

Research Associate Professor Ajmal Mian (UWA School of Computer Science and Software Engineering  (+61 8)  6488 2702
Michael Sinclair-Jones (UWA Public Affairs)  (+61 8)  6488 3229  /  (+61 4) 00 700 783