Less than 30 Seconds to Enroll
The system captures a user’s physical or behavioural characteristic, such as a fingerprint.
The software enhances the input from the sensor, removing the noise and unwanted elements to achieve the desired format for efficient feature extraction. Distinctive and repeatable features are selected.
Template Creation & Store
The cleansed and enhanced user data is stored in the feature template, which is stored in a database for later retrieval. Unwanted elements are removed to reduce the template’s size. The template is then used for matching in the future
Ezmcom’s biometric authentication was designed with two key criteria in mind: Security & Ease of use.Ease of use is a fairly straightforward concept; we wanted a fast and easy training procedure that would take under 20 seconds to enroll. We also wanted a fast access approach that could be done with one hand holding the phone and entrance into a device or application in a matter of a second. And we wanted to accomplish this without any unnatural or intrusive maneuvers. Finally, the goal was to give the right user the ability to get in without a lot of repeated attempts to enter.
Ease of use for Ezmcom was a major design criteria, and in our comparison assessment we gave it the highest rating, equal to that of face alone, which we believe is the least intrusive and easiest to use of all biometrics. We believe we have been successful in accomplishing our criteria for ease of use, and we welcome you to try out our technology, test it, and judge for yourself.The Ezmcom authentication system was initially designed to achieve excellent performance on existing public face databases such as Color FERET and YaleB. However, these databases are not very representative of the mobile device authentication challenge, so we also immediately began work on collecting a proprietary database of video and audio recorded directly from mobile devices using just the standard front-facing camera and microphone, across a wide range of environmental conditions. Our current primary dataset includes almost 200 individuals each with roughly 10-30 sessions recorded across a range of conditions.
The collection guidelines assume cooperative users who are attempting to authenticate under the expected range of environmental conditions in which a phone may be used.Ezmcom continues to use some public databases as test sets to investigate particular research concerns, but our algorithms are tuned to maximize accuracy for the mobile device application, accounting for the sensors and computing power available as well as the conditions likely to be encountered. We continue to expand our internal dataset, adding more subjects and sessions.Thus with a borderless IT landscape, the focus needs to shift from device based security to an application based security. One must also take note that while there is nothing like 100% security, a multi-layered security approach with new generation authentication including matured technologies like Voice Biometrics and Face Recognition can present a fighting chance for organisation against cyber attackers.
Our software is easily dropped into any
Appliance independent and requires no changes to networking infrastructure