ISJ Exclusive: Security, flexibility and unending possibilities
James Thorpe
Share this content
Experts at i-PRO explain why the company is leading the next generation of open platform AI for physical security.
On 1 October 2021, i-PRO EMEA was unleashed. Born out of the legacy of the Panasonic security brand and extending its well-known reliability, the new i-PRO entity brought something new to the market – a fresh agile approach that enables quick delivery times, stable prices and new levels of flexibility with openness on artificial intelligence AI-powered platforms.
The next-gen approach resonated and the company has quickly become a trusted global provider of surveillance systems specialising in edge-AI powered devices. Now, at our one-year anniversary, we’d like to share the specifics of our agile approach to AI edge-computing and lay out the potential and open platform AI approach brings to video security systems.
The leap forward in video surveillance
When it comes to machine and deep learning, the more precise the algorithms are, the more proactive the cameras equipped with image analysis can become when feeding information back to the user. The new generation of AI-enabled cameras allows us to go even further and make more precise pre-analyses of situations. In the case of a city, for example, new algorithms will be able to draw operators’ attention to crowds or abnormal crowd movements.
Deep learning algorithms, based on artificial neural networks, are inspired by the human brain and allow large volumes of data to be used to “train” an algorithm to recognise and categorise classes of objects. Once trained with all of this data, AI systems can recognise different shapes: A man, a vehicle, and also recognise colours; a man wearing a red T-shirt and a moustache, a child on a scooter, a particular face. Deep learning algorithms enable all of this as they’re capable of processing an exponential amount of data. Added advances in computing power and the miniaturisation of computer processors which are now integrated directly into security cameras at the edge, make the cost of deploying these technologies much more accessible.
Today’s video surveillance systems have become experts in pattern detection and recognition. Based on what they have detected, they create a metadatabase that is self-powered and allows for an even finer analysis of the images. This is also the basis of biometric and facial recognition tools, which must work hand in hand with the regulatory framework in Europe to strictly protect personal data and individual liberties (GDPR).
From a security perspective, the aim of metadata is not to identify citizens, but to optimise the use of video protection, making it more proactive and predictive. It enables connected in-car cameras that help drivers avoid accidents, smart cities that optimise traffic flows and limit incidents, even cameras that detect the micro-expressions of an individual having a stroke. The possibilities are unlimited and some that once sounded far-fetched are already a reality.
From their conception, modern security systems are designed to go beyond a protective logic. According to an Omdia research led in 2021, it is estimated that, by 2024, 30% of cameras sold on the market will be capable of embedding deep learning. AI-based image analysis is booming. To make these possibilities a reality is now only a matter of the time needed to deploy infrastructures and create the AI applications required to positively impact society.
AI at the edge
Cameras have already changed the way we live and work. Next, the combination of AI and the latest generation of open platform camera technology stands to change perceptions of this technology from video recording to data gathering. The industry is moving from reactive to proactive surveillance, able to intervene in an effort to prevent unfortunate events instead of just recording them. Doing so takes a burden off operators and frees their time.
In the past, Al-based video analysis typically took the form of a server-based system that performed arithmetic processing on video uploaded to a server. In recent years however, edge AI, which allows for deep learning arithmetic processing to be performed within each surveillance camera, is becoming widespread.
The main advantages of edge Al are that it reduces costs, provides better information security and enhances real time performance. Data is processed directly on the device that collects the data. Only the metadata, such as characteristics, behaviour patterns or attributes, are sent to the server or VMS. This approach takes a heavy load off the network and reduces latency enormously.
By installing various apps on an i-PRO Network Camera, organisations can easily adopt video analysis technology with edge Al. With the i-PRO Network Camera, both video and audio analyses are performed by edge Al processing.
AI and open-platform camera systems are truly a game changer. Imagine a world where the cameras are not just watching and reporting for security but are having a wider, positive impact on people’s lives – where cameras enable lights to switch on and off as people come and go, predict traffic jams and automatically re-route vehicles, open retail tills just before a queue starts to form and even identify customer shopping habits.
An important new aspect that i-PRO has brought to AI-driven cameras is making them open systems. This means there is a software developers kit (SDK) that enables developers to create their own purpose-built apps to run on the cameras.
i-PRO’s approach to privacy protection
As is true of many modern technologies, AI can initially be misunderstood and feared. As more AI cameras are deployed and used in accordance with GDPR regulations, the benefits are likely to assuage those fears.
Consider that the use of AI in cameras is about generating metadata for the detection and recognition of patterns, rather than identifying individual identities; i-PRO believes firmly that the continued safe use and unimagined benefits of AI and cameras can serve to shape our views of them – seeing them not just as a visual record of our lives, but as a means for enhancing them.
Seeking to understand and address privacy concerns, i-PRO has developed i-PRO AI Privacy Guard, an AI application that generates data without capturing individual identities, following European privacy regulations that are among the strictest in the world. i-PRO is also the only camera manufacturer that can run real time edge-AI privacy, split in four different modes pixelating or completely blocking either the face or the whole body of a person.
Managing metadata from edge devices
The most advanced next-gen cameras are capable of running four different AI applications at the same time. An important aspect of their capabilities is the accuracy by which they can classify objects and gather all the meta data required. Key to this accuracy is the number of attributes provided for segmentation. i-PRO AI-capable cameras classify 13 main attributes like gender, age and hair colour – as well as over 98 attribute items such as clothing colour and facial characteristics – to be able to properly classify data and deliver traceability of searched persons or objects.
i-PRO enables the highest number of attribute items on the edge and is the only manufacturer delivering up to eight directions of movement, which ensures proper follow-up for searches. i-PRO’s open platform approach works hand-in-hand with metadata to allow the capture and relay of data smoothly. This also ensures that searches are filtering out the correct information to deliver the desired outcome.
To enable operators to make decisions quickly and act proactively, i-PRO has developed i-PRO Active Guard, which processes AI data in real time. It is a key pillar of i-PRO’s overall AI strategy and is already integrated with Genetec and Milestone, with many additional partners to come. The plug-in is available free of charge, making it accessible to everyone at no extra cost.
There are many types of AI applications that can be installed on network cameras at no additional cost. Because i-PRO network cameras are edge devices, they manage various detection on the hardware itself. As a result, the camera is responsible for the advanced processing; the need of an expensive server is eliminated.
i-PRO Active Guard is a clear example of i-PRO’s newly implemented open policy and open platforms capability and it integrates seamlessly into existing infrastructure. The SDK allows developers to easily integrate third party applications, giving them access to a wide range of automation processes and applications.
AI-enabled remote monitoring
Remote monitoring is the monitoring of an entire security system, carried out by a third party monitoring provider. It typically involves a perimeter detection control system that has been designed to monitor sites such as factories and logistics hubs with the goal of deterring false alarms.
Remote monitoring operatives are trained to respond quickly and efficiently to alarm notifications. Remote security monitoring enables organisations to ensure that they act upon real alarm activations only, saving valuable time and money across the response chain. It is an additional service that will increase the overall safety and security of a site, whilst improving day-to-day efficiencies and business operations. Employees can focus on their roles and not worry about potential intruders entering the premises, especially out of hours.
Video analytics can enhance remote-monitored systems through visual verification; a foundation of the security industry. A connected sensor can detect the intrusion and send the information to the camera so the alarm can be visually verified and acted upon.
Utilising cameras with the latest AI-enabled edge analytics has the potential to revolutionise this process. By simplifying deployment and delivering more in terms of information with i-PRO AI cameras, ARC operatives can make informed judgements faster and more accurately and reduce false alarms, saving time and money.
A complete visual security environment
i-PRO is committed to bringing the benefits of AI-enabled video and audio analytics to the video surveillance market by including them in its major camera lines. We believe this technology is transformative for security professionals and the people and assets they protect. Recognising that having the best AI-enabled camera is only half of the solution and we created our Active Guard plugin to bring this wealth of additional data to major VMS.
While having additional object and attribute data improves search and event notification performance, it doesn’t have to do so by sacrificing privacy. i-PRO’s Privacy Guard technology ensures that data collected is anonymous and compliant with GDPR conventions. We believe that system integrators and end users should be able to choose from a variety of best-of-breed analytic applications that best suit their unique requirements.
The i-PRO open AI platform enables developers to design purpose-built applications that can be seamlessly installed in our cameras, providing a level of customisation and flexibility unique in the industry.
To find out more information, visit: https://i-pro.com/
This article was originally published in the October edition of International Security Journal. To read your FREE digital edition, click here.