Face-Off: Top Identity Verification Systems Compared

In the era of rapidly advancing technology, the development of a new face identification test is of paramount importance, particularly in light of the growing use of facial recognition software in police investigations and security settings.

Psychologists from UNSW Sydney have developed the Glasgow Face Matching Test 2 (GFMT2), a test designed to target high-performing facial recognition individuals, commonly known as super-recognisers. This test was developed in collaboration with the University of York and the Australian Passport Office.

The new GFMT2 test is a response to the requirements of professional groups and is intended for roles such as visa processors, passport issuers, border control officers, police, contract tracers, security staff in private industry, and others.

The test involves matching faces across changes that make face identification difficult in real-world tasks. Scientific understanding of face perception has made tremendous advances in the last ten years, and this update gave the developers a chance to build in new knowledge.

The GFMT2 test was published in the journal Behavior Research Methods, and the DOI for the publication is 10.3758/s13428-021-01638-x. All tests are available free for scientific use.

Recent work suggests that current professional training courses in face identification do not improve people's performance. The improved GFMT2 test aims to address this issue by allowing before and after comparisons to assess the impact of training and other interventions.

The development of the GFMT2 test was part of an ARC Linkage Project in collaboration between UNSW Sydney and the Australian Passport Office (DFAT). One of the key benefits of the GFMT2 test is that super-recognisers revealed by the test can be used to review identity matches made by facial recognition software and help prevent false identity matches.

The proposed introduction of a centralised database by the Australian federal government, which would host photos collected by a range of state and federal agencies, has been met with criticism due to concerns about errors in face recognition software leading to wrongful arrests. The GFMT2 test could potentially play a significant role in mitigating these concerns.

At present, there are no publicly available or specific search results identifying Australian authorities or institutions that intend to use the new version of the Glasgow Face Matching Test (GFMT2) for person identification with facial recognition software. However, the test will be able to be used in a number of countries including Australia, New Zealand, UK, and Canada.

In conclusion, the GFMT2 test is a valuable tool in the ongoing effort to improve the reliability and accuracy of facial recognition software. By identifying super-recognisers and allowing for before and after comparisons, the test could significantly enhance the performance of facial recognition software and help prevent false identity matches.