A recent innovative launch in the multimodal biometrics space is Armatura’s FAPV30 multimodal biometric series. These include AMT-FAPVS-30, a built-in standalone multimodal recognition module for face and palm that can embedded into the host device; and AMT-FAPVR-30, an external standalone multimodal recognition reader for face and palm.
In the field of access control and identity management, biometrics have become an important means of authentication. In particular, more and more attention has been focused on multimodal biometrics, which combines more than one biometric technology into a solution.
A recent innovative launch in this space is Armatura’s FAPV30 multimodal biometric series. These include AMT-FAPVS-30, a built-in standalone multimodal recognition module for face and palm that can embedded into the host device; and AMT-FAPVR-30, an external standalone multimodal recognition reader for face and palm. Both comes in compact sizes that fits a wide range of host devices without needing modification.
Sharing the same PCB board, both low power consumption biometrics devices can be integrated directly into third-party brand readers connected by the USB 2.0 interface. This makes the FAPV30 models suitable for any system/application – including time attendance, access control, point of sales and visitor management – where physical access authentication is required.
“The FAPVS/R-30 series products employ multimodal biometric technology but it does not mean that both face and palm are always combined together for one operation. Instead, it provides very flexible options,” said Simon Zhang, Product Manager for Armatura. “The user can either choose face mode, palm mode or combined for recognition depending on their business requirements and local regulation policies. Of course, the combination mode approach with face and palm provides the highest security because of its excellent performance, with the total time for the clock-in process completed within one second.”
Existing use cases can be found across Asia, North America and Europe – a US manufacturer has chosen the FAPVS-30 module for their commercial intercoms, while a Korean security company installed the module inside a standalone access control device to help them identify employees via facial recognition. A Chilean casino venue used the standalone FAPVR-30 reader in its auto redemption machine for employee verification by facial recognition. Additionally, a few companies have been testing a pay-by-palm e-payment system using Armatura’s palm vein technology.
Armatura FAPV30 multimodal biometric series
Equipped for performance
Facial recognition has greatly matured over the past few years due to continuous advancements in deep learning technology, offering users both accuracy and a user-friendly experience. Adding palm vein and palm print as an authentication factor only increases the overall security of the system as this biometric trait can only be actively collected – meaning that user palm images can only be taken intentionally: with consent and specialized cameras – unlike facial images that can be taken from the internet and without permission. This dual biometric approach adds an extra layer of security and privacy to the security system.
These types of systems, however, require higher processing power and larger datasets. The Armatura FAPVS-30 and FAPVR-30 are equipped with a high-performance quad-core ARM CPU @1.2GHZ with NPU architecture, providing 2.4 TOPS computing power to run deep learning algorithms. The module features a 1.0 GB RAM and 512 MB storage space for large-volume face and palm templates (30,000 faces and 5,000 palms) for rapid and accurate database searches during biometric authentication.
When it comes to speed, Zhang states that the built-in biometric algorithms on the FAPV30 series can complete one face recognition in less than 100 milliseconds or 10 faces per second, while palm recognition takes less than 140 milliseconds or 7 palms per second.
Built for accuracy
High-definition images are necessary for higher reader accuracy. The Armatura FAPV30 multimodal biometric series uses 2MP RGB and 1.3MP NIR cameras to capture high-resolution images of the face, palm and even palm vein patterns. The cameras also provide RGB and NIR image-based liveness detection that has been effective against spoofing attacks to help keep the system secure.
With regards to the accuracy and efficiency of its independently-developed algorithm, earlier this year Armatura reported that it was a
top performer in the NIST Face Recognition Vendor Test (FRVT), an industry benchmark when it comes to evaluate facial recognition algorithms for verification (1:1) and identification (1:N)
Developer-friendly interfaces
The Armatura FAPV30 multimodal biometric series is supported by the MultiBio SDK 3.0, which offers developer-friendly interfaces for machine-to-machine communication over UVC (USB video class) and HID (Human Interface Device) protocols, and supports Windows and Android operating systems along with Linux system which is on request.
With this (and detailed hardware integration and installer user manuals), a software developer with limited to no hardware experience can set up the face/palm matching SDKs to run on the host device (with the modules only extracting the template) and allow user templates to be managed and authenticated directly on the host application. Users can access and choose from a range of biometric functionalities to fit their needs, such as face/palm detection from video frames, liveness detection, facial analysis on gender identification, age estimation, expression classification, glasses/hat detection, as well as face/palm feature extraction and matching.
This development lowers the biometric development bar for non-biometric developers and streamlines FAPV30 integration development to bring products to the market faster.
Conclusion
Multimodal biometrics have gained much traction in access control due to their increased security and price-performance ratio. Against this backdrop, Armatura has introduced the multimodal FAPVS-30 module and FAPVR-30 reader with hardware and software features that fully showcase the company’s manufacturing, development and integration capabilities.