On September 11 2001, a terrorist attack on America, one of the most powerful countries in the world, brutally highlighted the need for increased security in the new millennium. The plethora of dangerous locations that need security such as airports, train stations and chemical plants to name a few, makes security a very lucrative research field. Because of the enormous monetary benefits the field was already littered with recognition methods prior to September 11. Iris scans, finger and palm print scans, retina scans and DNA matching are such techniques. These methods are very accurate (99.9999% in some cases) and are very hard to deceive. Yet September 11 still occurred even though two of the pilots were on FBI watch lists and were captured in surveillance footage.
Face scanning is not so accurate and is much easier to deceive than the other biometric methods already mentioned, but because it can be done covertly at a great distance from an uncooperative subject, it has its own niche applications, such as making the connection between surveillance footage and the person being on the FBI watch list, automatically. While other methods such as iris scans have their place in private institutions such as vault access in a bank, non-obtrusive biometrics such as face scanning are ideal for un-secure public galleries or even for probing the streets for known offenders as an example.
My research aim is to create/research an application that can automatically use one or more web cameras to recognise faces in image captures, from a database of 3D faces. This task must be accomplished faster; more accurately and more robustly than current (at the time of thesis submission) facial biometric technologies manage. My algorithms will make use of neural networks, Bayesian networks and various image processing techniques.
About the speaker:
Mr. Karl Axnick
Course: Research in Master of Engineering and Science in the field of Computer Vision
Supervisors: Professor Raymond Jarvis
Associate Professor Kim C Ng
Previous degree: Graduate from Monash University with Bachelor of Electrical and Computer Systems Engineering degree.