printer friendly version

Unleash the full potential of AI at the edge

Edge devices are commonly used in access control systems because they are easy to install, scalable, and cost-efficient.

In the context of access control, edge devices are located near access points, such as doors or gates, to process authentication, manage access permissions, and make

real-time decisions at the point of entry. Edge devices are often preferred over traditional access control systems, where the permanent presence of a person is required to grant access. Edge devices are cost-effective, easy to install anywhere, and can scale up based on the organisation’s needs. Edge devices are widely used in various sites, from small and medium-sized businesses to enterprises where real-time 24/7 access control service is essential.

However, edge devices are small, low-power, and low-cost. As AI (artificial intelligence) adoption increases in the security industry, edge devices require more advanced hardware and software technologies to support the deployment of AI algorithms. Edge AI refers to deploying AI algorithms and running AI models directly on local edge devices. Therefore, optimising and streamlining AI algorithms for implementation on edge devices is crucial due to their limited resources, such as memory, storage, and processing power.

Running and training AI algorithms require vast datasets and high hardware specifications, as AI algorithms need to learn from large amounts of data and perform complex computations.

Feel the power of Suprema’s Edge AI, which runs NPU (Neural Processing Unit)-based AI algorithms on edge devices locally.

By integrating AI and edge computing, Suprema has developed high-performance edge AI devices. Suprema has adopted an advanced AI processor, NPU. The NPU is designed to execute AI algorithms with higher efficiency and performance compared to traditional GPU (Graphics Processing Unit). The NPU can process AI workloads faster while consuming less power, making it ideal for applications in edge devices and the security industry where real-time processing is crucial.

A standout feature of Suprema’s edge AI technology is its ability to create lightweight AI engines capable of efficiently handling vast datasets, while preventing overheating and slowdowns in hardware. Suprema’s optimised data processing model, known as the quantisation formula, plays a pivotal role in reducing both model size and operational quantity. This not only enhances model efficiency, but also bolsters hardware performance. The fingerprint recognition solution, BioStation 2a, and facial authentication solution, BioStation 3, can run AI algorithms and process authentication locally.

What sets Suprema apart is its capability to develop deep learning-based AI algorithms for biometric recognition.

AI integration proves particularly advantageous in biometric-based access control systems, where it analyses biometric data with higher precision. AI can identify distinctive patterns and features that might be difficult for traditional algorithms to discern, leading to improved recognition accuracy.

Additionally, AI can incorporate feedback from past errors to improve its algorithms, leading to better overall performance. Leveraging over two decades of extensive experience in the biometrics and access control industry, Suprema has the capacity to train on vast datasets for deep learning. Suprema stands out for its technical superiority in offering the most accurate and fast biometric recognition performance through data-driven, deep learning-based AI algorithms.

Share this article:

Further reading: