In 2009-2010, IBG was awarded a series of four research grants through Defence Research and Development Canada’s Public Safety Technology Program. These research grants illustrate IBG’s capabilities in advanced R&D in challenging biometric areas. The following Case Study summarizes IBG’s Study Report on stand-off face and iris recognition systems.

IBG performed a comparative analysis of commercially-available stand-off biometric systems, including dedicated iris recognition systems and combined face and iris systems from vendors including Sarnoff, AOptix, Hoyos, Honeywell, 3VR, and Cross Match. IBG evaluated features, benefits, and differentiators, describing factors that may inhibit collection and matching performance.

IBG developed Subject Acquisition Profiles (SAPs) to describe collection, matching, and imaging, and other salient requirements for stand-off systems. SAPs define parameters of interest for biometric systems (e.g. capture volume) and specify target specifications or performance requirements for each parameter (e.g. ability to capture a subject moving at 1m per second). SAPs simplify procurement by using a numerical scale that summarizes a devices’ collective feature set. SAPs also directly or indirectly support interoperability across different vendor systems. The SAP model has been used successfully in the biometric industry to categorize different types of fingerprints devices and mobile biometric devices.

To evaluate the performance of biometric systems in an application with a moderate stand-off distance between Test Subjects and capture devices, the team conducted a field study of two stand-off systems: (1) the AOptix InSight iris recognition system and (2) a custom-designed face recognition system using Neurotechnology VeriLook 4.0. Field Study results illustrate the viability of stand-off iris recognition in applications with non-habituated Test Subjects.

Having testing dozens of iris recognition systems since the 1990s, IBG observes that complexity of subject-device interactions has been a primary impediment iris recognition adoption. In particular, unattended operation by non-habituated subjects has been a challenge. Empirical and informal AOptix InSight results underscore the device’s simplicity of interaction and its ability to tolerate variability in subject positioning. The Field Study also illustrated that even under non-ideal conditions and without any operational optimization, face recognition can provide reasonable identification performance on a closed set of test subjects, as would be encountered in an access control application.