Multi-enrolment as a way to speed up processing time
Aurora-AI is best known for development of Artificial Intelligence solutions. One example being our world-class Infrared (IR) Face Recognition systems, with thousands of installations globally including major implementations at Heathrow and Manchester, which have been helping to secure the UK Border for many years. The use of IR provides stability against changes in lighting conditions, meaning our performance is consistent year-round with no impact from sunlight, dull lighting or shadow casting.
The IR capture technology improves every year. The first-generation cameras streamed at 4 frames per second, had a fixed lens configuration and no polarisation; the latest generation can stream at over 12 FPS, has variable lens/filter configuration, higher resolution and most importantly supports HDR (High Dynamic Range). The HDR feature is essentially a technology that assigns 12 bits per pixel, allowing for 4096 pixel intensity values in contrast to the standard cameras providing only 8 bits per pixel for 256 values. The extra information means that 8 bit images can be reconstructed out of 12 bits using clever bit shifting and compression techniques. The new HDR 12-bit information has been utilised by our research team to improve the speed of the face detection by implementing a specially designed 12bit face detector. The faster frame rate and the ability to detect faces in HDR means that people can be detected while approaching the camera, at greater distances, even when the scene is very dark. This leads to faster identification, even before the subject has reached the camera. The HDR information is also used in the face brightness adjustment algorithm, allowing creation of perfectly illuminated images of the face even when the IR flash is not fully adjusted during periods of rapid movement.
The HDR algorithm adjusts the face illumination independently for every face seen on the image using a single flash power setting. The face brightness is controlled by our software and used to guarantee perfect illumination independently of the environmental factors like proximity or intensity of other light sources. The flash adjustment is an integral part of our face recognition technology and is being controlled on all our cameras. What’s more, this technique of independent face-region adjustment then opens the possibility of multi-subject enrolment as explained below.
The multi-face flash adjustment technology (available only on HDR camera models) is the next generation of our infrared brightness control functionality, allowing us to use a single camera and create perfectly illuminated images of every face, that can later be used in biometric transactions.
One such biometric feature is multi-people enrolment. The functionality takes images of people visible to the camera and groups them together across multiple camera frames. The grouping algorithm uses the location as well as face-match comparison scores generated by our Artificial Intelligence biometric engine, creating unique enrolment templates per person. The entire process is executed as a single transaction and results in multiple distinct enrolments created automatically with no need for manual human intervention. The templates can later be used in long and short-range identifications and verifications.
One of the most common use cases for the multi–enrolment functionality is simultaneous enrolment/verification of families. This means that a whole family group can enrol in a single step, saving huge amounts of time and mitigating against some of these exceptional cases, which would otherwise be treated outside of the usual process. The use case has been described by Heathrow staff as a “game changer” for biometric implementation in airport processing.
To see this exciting new technology in action, please visit us at PTE 2019 in London ExCel (26th – 28th March). You’ll find us on stand #1090. Please use our contact page to make an appointment.