Advances in processor design and fabrication have made artificial intelligence (AI) processing acceleration more widely available, democratizing the technology. While mobile device processor makers started adding AI accelerators to their system on chip (SoC) designs several years ago, edge devices are only now benefiting from on-device access to affordable, efficient, yet powerful AI acceleration.
Advances in video surveillance are being enabled by efficient new processors like the Axis ARTPEC-8 SoC. Axis Communications has started rolling out network cameras with these AI acceleration processors inside and has plans to make similar enhancements to its broad range of surveillance products.
Furthermore, applications like CyberLink’s FaceMe Security can make the most of this newfound on-device processing power. Thanks to the recently added support for FaceMe Security in the Axis Camera Application Platform (ACAP), and FaceMe SDK's support of the Axis ARTPEC-8 SoC, customers can see big improvements in smart video management system functionality, with faster performance, and significantly expanded video coverage.
Read on to learn what newly forged partnership between CyberLink FaceMe and Axis Communications, means to you as a video surveillance systems integrator or end user.
Axis Communications is a leading provider of video surveillance, access controls, intercoms, and audio systems. The Axis Camera Application Platform (ACAP) is an open application software platform adaptable for a broad range of industries and use cases. You could think of it as a hub application, where you can configure, access, and use a wide selection of networked cameras within a range of data driven surveillance apps.
Thanks to its built-in flexibility, ACAP can be used to create customized applications to fit the specific needs of a broad range of industries. These industries, with characteristically different operations and priorities, naturally benefit from different ways of accessing surveillance hardware, and analyzing gathered data. This means that ACAP is great for both system integrators and their customers, as it facilitates finely customized applications.
Axis continually develops ACAP to keep up with state-of-the-art hardware, development tools, and infrastructure. The goal is to future-proof ACAP solutions that are in place, enabling users to adopt the latest hardware plus computer vision and video analytics software innovations. Despite the 'Camera' in its name, ACAP can also integrate feeds and data from other networked Axis security hardware like speakers, intercoms and radar products.
One of the ways that Axis makes ACAP one of the best solutions for customers is in allowing the integration of technology from partner companies, like CyberLink. By joining the Axis Technology Integration Partner Program, users of supported cameras, like Axis IP cameras, can now leverage FaceMe's leading technology for tasks like identity authentication, attendance management, access control, security alerts, and AI smart monitoring.
The tight integration of FaceMe within ACAP means that Axis network cameras, with built-in AI-accelerating SoCs, can shoulder much of the AI processing burden for FaceMe, enabling faster processing and / or simultaneous monitoring of more video channels.
ACAP flexibly combines advanced edge analytics and cloud or server-based technologies to run applications completely or partially at the edge. In this article we will look closely at edge use cases enabled by Axis's own custom processor, the SIA (Security Industry Association) award winning ARTPEC-8 SoC.
As its name implies, the ARTPEC-8 is the eighth generation of this custom SoC, designed specifically for accelerated network video analysis. The chip delivers deep learning and AI analytics on the edge and is built-in to the newest Axis cameras. This edge processing, combined with cameras packing the newest image sensors, yields the fastest and most accurate analytics and alerts.
The FaceMe SDK is a leading, highly accurate, AI facial recognition engine, purposely designed to integrate edge-based AI facial recognition into a wide range of IoT and AIoT solutions. FaceMe's SDK supports a wide range of platforms (Windows, Linux, Android, iOS and more), architectures, AI inference engines, and GPU / VPU hardware acceleration choices - including the latest ARTPEC-8 SoC from Axis.
For time-sensitive tasks where the most accurate AI-based facial recognition technology is deployed, fast edge processing is desirable. Situations where solutions like FaceMe get deployed will often benefit from rapid data processing and analysis, for the timeliest results. Cloud processing options may be a welcome alternative or backup choice but can be prone to adverse network conditions and latency, as well as other drawbacks. Another choice, using a server situated near the edge devices adds complexity, as well as hardware, maintenance and other undesirable costs.
In our 2023 Ultimate Guide for Facial Recognition Technology a more thorough appraisal of edge-based facial recognition is provided. In brief, as well as the processing / response speed benefits compared to cloud processing, outlined above, organizations may welcome: cost of ownership benefits, greater service reliability and availability, install flexibility, and stronger security.
Axis has designed its ACAP cameras to make edge processing capable camera installations simple and incredibly effective. With an ARTPEC-8 SoC in each camera, the processing power is available to analyze data on the spot. For example, the Axis SoC can run real-time object detection and classification on anything that enters its field of vision. Additionally, thanks to FaceMe SDK integration, Axis cameras like the Axis Q3536-LVE/38-LVE Dome Cameras and Axis Q1656-LE Box Camera can run AI-enhanced facial recognition on-device. Remember, FaceMe is the world's top ranked NIST FRVT 1:1 and 1:N facial recognition solution with 99.83% accuracy, anti-spoofing protected verification. Being able to run this software entirely on the edge, thanks to a tiny chip built into a camera is no small feat.
If you’re unsure whether edge or cloud-based processing is for you, or you’re thinking about a hybrid solution, you could still find ACAP appealing. The platform combines advanced edge analytics and server-based technologies using your choice of cloud, local server, and edge environments; allowing you to choose to run applications completely or partially on the edge. Not only does ACAP provide obvious flexibility, this hybrid ability can also improve system efficiency and reduce storage, bandwidth, and hardware needs, for simpler, scalable and efficient deployments.
Axis currently has a basic computer vision solution available to ACAP users with connected cameras, in the form of its Axis Face Detector app. However, it is limited to recording and detecting shoplifters, or similar block-listing style functionality.
Benefiting from the superlative flexibility, accuracy, and performance of CyberLink's AI facial recognition engine, FaceMe, and the latest Axis cameras is relatively simple. A developer only needs to create an app which installs within ACAP using the free SDKs from Axis. Then, as FaceMe Security is certified compatible with ACAP, users can benefit from an upgraded smart security and access control infrastructure. Specifically, FaceMe can be used to provide advanced and automated computer vision based processes like: identity authentication, attendance management, access control, security alerts, and AI smart monitoring for ACAP camera users.
Some developers and their clients may only require light-weight video analytics at the edge, while others may favor taking full advantage of edge hardware acceleration as provided by processors like the Axis ARTPEC-8 SoC and cameras which have it built-in, to run advanced facial tracking and face recognition on the edge.
A video management system (VMS) can be considered the heart of a complete video surveillance operation. In the most basic typical monitoring system there are strategically positioned cameras, which in turn are connected to a video recording device and a bank of video displays. Before the era of affordable, accessible, and powerful computer processing, the VMS might be used by security employees for monitoring entirely by eye. Their only aids would be video display controls like switching to a larger screen, and using transport controls like play, pause, fast forward and reverse to review footage. These controls went digital with the dawn of computer video and mass storage, but the biggest revolution occurred much more recently with the dawn of computer vision and AI-based facial recognition. VMS remain an essential component of video surveillance.
Axis has its own popular VMS, the Axis Camera Station. This open platform is the most natural fit for software-based solutions written for Axis branded devices. Using Axis Camera Station as your VMS with an Axis-based system seems sensible as it is a free tool for use with Axis hardware, thereby helping keep a surveillance system's TCO (total cost of ownership) low.
The Axis Camera Station is billed by its designer as powerful, yet easy to operate. It offers an intuitive user interface (UI) so that personnel can quickly master the basics and explore the depth of functionality. Out-of-the-box this VMS enables: video surveillance, security and access control, search functionality, and privacy features, complete with a mobile viewing app for remote monitoring. Despite its name, the Axis Camera Station is not just compatible with Axis cameras, it also integrates support for related Axis hardware like intercoms, AV recorders, wearables, and more.
Axis Camera Station is therefore a good app to get started, particularly when paired with various Axis branded hardware, as the Axis product's usefulness via the integration of advanced AI-infused technologies can deliver a lot of benefits.
Photo courtesy of Axis Communications AB. All rights reserved.
The integration of FaceMe Security with the Axis Camera Station VMS provides functionality founded upon CyberLink's highly-ranked and precise FaceMe AI facial recognition engine right within the VMS UI. Previously, the Axis Camera Station required trained security personnel monitoring for anything but the most basic surveillance situations.
Axis provides support for its VMS being used for security and access control based on personnel carrying security cards, keys or tokens. Facial recognition strengthens security by shifting the identity verification paradigm away from something a person has in their possession to their true identity - their face. AI-powered identity verification's inherent advantage is that it can't be forgotten, lost, stolen or borrowed. Moreover, it can run with unrivalled accuracy and speed 24/7, 365 days a year.
The Axis Camera Application Platform (ACAP) enables FaceMe Security on its Axis ARTPEC-8 SoC powered cameras, so FaceMe’s facial recognition can now be done at the edge, running directly on Axis cameras. As a result, there is a reduced workload on the facial recognition engine, for improved performance and the capability to monitor more simultaneous video channels.
In addition to improved performance, the ACAP integration means attractive FaceMe features like person-of-interest detection, group tagging and management, and visitor summaries can be seamlessly layered on top of existing video feeds.
The collaboration between Axis Communications and CyberLink FaceMe developers has resulted in the most advanced AI-based facial recognition algorithms being able to run on key commercial security hardware like cameras, at the network edge.
Solutions arising from this partnership will provide added value to customers, measurable in terms of greater efficiencies and capabilities. Specifically, the advances come from integrating continuously improving technologies like the FaceMe facial recognition SDK, FaceMe Security, the Axis Camera Station VMS, ACAP, the Axis ARTPEC-8 SoC, and the newest network cameras from Axis.