Face is one of the easiest ways to distinguish one person from another. Even a toddler can recognise the mother’s face several days after birth. It is more intuitive to recognise a person by face than by other factors such as their walk, voice, and height.
Biometric authentication systems use physical traits such as fingerprint, face, iris and vein as a credential. Among others, facial authentication terminals use a face as a credential, which is the most familiar way for a human to recognise another person. Due to the nature of facial authentication technology, people feel comfortable performing facial authentication.
Fingerprint recognition devices, which are the most commonly used for biometric authentication, are equipped with an optical sensor. Facial authentication uses cameras to identify a person, so it does not require an optical sensor. Thus, it allows a user to perform authentication without physical contact. Since facial authentication terminals use IR (Infrared) technology, one of the major drawbacks is the limitation of installable location. A facial authentication terminal will perform poorly when it is installed outside or near windows because of the strong ambient light.
FaceStation 2, Suprema’s newest facial authentication terminal, goes beyond this limitation. It is an access control, time and attendance terminal, featuring a better user experience with Android 5.0 Lollipop and Suprema’s latest algorithm, hardware, and software. A wide range of functions and enhanced performance will provide users with a new non-contact biometric authentication experience.
Face authentication and near-infrared
Processing ambient light is critical to facial authentication performance. Most facial authentication terminals in the market have limited installable locations because the recognition performance varies depends on the strength of ambient light.
FaceStation 2 has 80 wide-angle near-infrared LEDs and 60 narrow-angle near-infrared LEDs, so it can recognise a face even in an environment with 25 000 lux, which is equivalent to an environment with full daylight (not direct sun). This allows users to install the terminals at indoor locations near windows, lobbies and building entries.
Another technology adopted by FaceStation 2 to enhance the performance is pixel intensity distribution analysis. One of the difficulties in using near-infrared LED technology is the effects of ambient lighting. Ambient lighting can make near-infrared LED lighting useless since it contains ultraviolet rays. Plus, the shadow on a face cast by ambient light can make extracting facial features difficult for the face feature extraction algorithm.
The three-dimensional pixel intensity distribution analysis minimises the effects of ambient light when acquiring facial images. As a result, the terminal acquires near-infrared images with minimal variation of contrast. It is easier for the algorithm to recognise the shape of the face with these uniform contrast images than with too bright or dark images, so it can extract more varied features, thus creating high-quality face templates. High-quality face templates are crucial for the performance of facial authentication.
The angle and position of the cameras on FaceStation 2 are determined so that tall users do not need to bend their backs too much. Plus, the different angles of view between the built-in visual and infrared cameras allow users to stand at a position that is most suitable for facial authentication (users can easily locate themselves by trying to locate their face at the correct position on the screen).
Authentication performance with 1.4 GHz quad-core processor
FaceStation 2 demonstrates exceptional performance, using its enhanced face template extraction technology and 1.4 GHz quad-core processor. It can perform up to 3000 matchings (1:N) within one second. The Group Matching feature can perform up to 30 000 matchings within one second.
Accommodating multi-RFID technologies
FaceStation 2 features multi-RFID technology, supporting 125 kHz and 13.56 MHz cards. If a customer is using 125 kHz cards such as EM and HID Prox and wants to change the access cards to more advanced and secure 13.56 MHz cards (Mifare, Mifare DESFire, Felica, iClass), then employees at the customer’s site can use both 125 kHz and 13.56 MHz during the transition period. FaceStation 2 also supports NFC and BLE communication technologies.
SIP-based VoIP videophone
FaceStation 2 also works as a SIP (session initiation protocol) videophone, eliminating the need to install a separate videophone. If a SIP server is already installed at the site, you can use the existing SIP infrastructure. Otherwise, customers can install a well-known open-source SIP server recommended and tested by Suprema to use FaceStation 2 as a videophone.
Live Face Detection
FaceStation 2 analyses the actual photo and the IR image based on its dual camera system and intelligent machine learning algorithm. Hardware-dependent live face detection systems (e.g., facial thermogram recognition, facial vein recognition) require expensive hardware components and provide less accurate matches and slower performance in authentication. Suprema’s live face detection technology gives users ideal accuracy and speed for access control and time and attendance management.
|+27 11 784 3952
|086 552 0819
|More information and articles about Suprema
© Technews Publishing (Pty) Ltd. | All Rights Reserved.