Facial Recognition extended to retail payment

Facial recognition is among one of the hottest technology trends nowadays and the booming applications of it is clearly foreseeable, starting from security-critical to retail payment systems. However, like other IoT applications, facial recognition generates tremendous volume of data, which may lead to performance disappointment due to network congestions. Therefore, with the assistance of Multi-access Edge Computing (MEC) for the anticipated 5G network specifications, the concept of facial recognition to be deployed at retail setting has become a reality.

Conventionally, facial recognition is deployed at enterprises and mission-critical environments where restricted access is imposed. But today, as 5G and AI (Artificial Intelligence) are around the corner, this technology can be applied to offer interactive experiences for consumers at a shop to make their payments by just scanning their faces, and this is foreseen as smart retail which requires high-authentication/identification speed contributed by the edge networking architecture – Multi-access Edge Computing (MEC).

How Multiple-access Edge Computing (MEC) Improves Performance

Since 5G indicates the shift of computing and analysis from the cloud to the edge, MEC(Multiple-access Edge Computing) could symbolize a step further for the edge computing. With multiple types of accesses at the edge, this network architecture allows more efficient real-time computing, higher bandwidth, and lower latency as MEC help reduce the congestion and response time in centralized cloud model.

MEC grants edge network hardware the ability to process, analyze, store and deliver data generated by users, so that the data does not have to travel to cloud. Thus, MEC model reduces the transfer volume by the backhaul of mobile broadband and other client devices.

The primary advantage of MEC is that the content host can deploy their network infrastructure closer to the users, and thus the performance in application performance and content delivery is greatly improved due to shortened response time and lowered latency. Today, MEC has helped enterprises achieve applications such as AR (Augmented Reality) and VR (Virtual Reality), connected cars and facial recognitions where large volume of data is generated on a daily basis, since MEC frees up the bandwidth and traffic that may be heavily occupied in centralized cloud architecture.

When deploying facial recognition in smart retail, the critical part lies in the payment authentication process and security. First, the completion of authentication for payment must be minimized in order to retain user satisfaction. On the other hand, Facial recognition relies heavily on IP cameras and also how fast the digital images can be processed. The system must be able to minimize the time needed to process the captured images and perform analysis, in order to enhance the real-time application performance. Secondly, the payment system involves the users’ facial and credit card information. The system must be able to protect this type of sensitive and critical data from hackers or system error that may cause leakage.

Recommended MEC Server Hardware

To optimize 5G and MEC application performance and user experience, service providers must enhance their edge network infrastructures to ensure higher bandwidth, lower latency and improved performance, so that intelligent edge computing can be realized. For instance, white-box hardware servers empowered by Intel® Xeon® D-2100 processor, like Lanner’s NCA-4020 and NCA-1611, are the optimal fit for edge applications.

Both NCA-4020 and NCA-1611 are driven Intel® Xeon® D-2100 processor architecture, delivering ultra performance for advanced analytics and offering scalability for network edge, while power consumption is minimized. Besides, there are hardware-assisted security mechanisms to protect the information generated at the endpoint devices, such as facial recognition in retail environments.

Fetured Products


NCA-4020

Next-gen vCPE/uCPE Platform for Accelerated SD-WAN Deployment

CPU Intel® Xeon® D2100 4~16 Cores
Chipset None

Read more
 

NCA-1611

NEBS-compliant Desktop vCPE Appliance for SD-WAN and SD-Security

CPU Intel® Xeon® D-1500 Series 4~8 Cores (Broadwell-DE NS)
Chipset SoC

Read more