Improving security solutions with AI and edge computing

Thanks to publicly available tools like ChatGPT, artificial intelligence (AI) has become an incredibly hot topic lately. AI has the potential to revolutionize many aspects of our modern world, and security solutions are no exception. AI could fundamentally change the way we think about security as we observe its applications in facial recognition, anomaly detection, and predictive analytics. Security solutions powered by AI feature faster response times and more accurate threat detection, making them valuable across multiple industries.

As more cameras are deployed around the world to monitor public and private property, there is an increased need to leverage artificial intelligence (or “AI”) to leverage the metadata in video streams that can be leveraged when using intelligent IP video products . These IoT systems generate huge amounts of data every second and every day for a variety of reasons. For surveillance devices, this data can be used from security, comfort or emergency situations. Whether they’re collecting data for a business to improve its operations or providing alerts about a security issue, cameras are evolving into proactive business and security tools, rather than simply providing forensic or reactive investigative capabilities.

Much of this surveillance data has potential value, but must first be transmitted, processed, stored, and analyzed. In the currently most common model, all data from the connected device is transferred to a data center or server for storage and analysis. Because not all data is useful or valuable, its transmission and storage can result in significant wastage of bandwidth and storage resources, not to mention the upfront energy consumption and cost of housing central processing power. Enter edge computing.

Edge computing brings increased computing power to the “edge” of the network, or more specifically, into the network video camera itself. This allows for some level of data analysis by the device, and therefore the transmission of only meaningful, useful data, or data awaiting further analysis (e.g. to alert officials to exceptions at border control where passport verification is required). The benefits in terms of bandwidth and storage requirements are obvious, not to mention those that come with increased operational efficiency. Since data transmission often requires compression, edge computing can bypass this and, combined with AI-based analytics, provide the clearest possible picture for the AI ​​to work on.

READ :  Spectro Cloud Palette Edge enhancements boost security in edge Kubernetes environments

meaning of the edge

The “edge” of a network refers to the computing infrastructure closer to the data sources within a given system. In an AI-based security solution, these are typically the video surveillance devices themselves, which over the years have evolved into increasingly powerful and powerful network computing devices that use high quality lenses as the primary sensors to collect data. Edge computing can be an important aspect of AI-based security solutions as it reduces latency through proximity and helps keep sensitive data local, which can reduce the risk of data breaches. For example, edge solutions can function without a server or cloud-based connection, meaning a single malicious actor cannot gain access to an entire system by affecting one aspect of its operations.

The added benefit of running analytics at the edge, especially when it comes to cameras and analytics, is the ability to run analytics on an uncompressed image, resulting in higher fidelity and metadata. With traditional server-based deployments, the camera first compresses the video stream to save bandwidth and the server-based analytics are performed on the compressed stream.

Improving security solutions with AI

There are several ways to use AI to increase the security of a given facility. Face recognition, anomaly detection, and predictive analytics all contribute to a safer environment that can respond more quickly to threats. At the highest level, AI-based security solutions detect threats faster and more accurately; This has numerous practical applications, such as detecting fraudulent financial transactions or identifying illegal items in security X-rays.

For example, facial recognition technology can be used to identify people entering a secure area, or to assess a group of people and identify people who may pose a security risk. In the United States, the Transportation Security Administration (TSA) has tested AI-based facial recognition to scan passengers’ faces and compare them to a database of known security threats. Lower-level solutions come in the form of “face detector” software, which aims to deter thieves by giving the illusion that they’re being followed in a retail store by sharing an audio message letting passers-by know they’re being watched become .

READ :  Intel sales collapse in 2022, expect no recovery in Q1

Of course, we can’t talk about AI and facial recognition without addressing privacy and restrictions on the collection of personally identifiable information (PII). Options include static privacy masking solutions, ideal for indoor or outdoor scenes with fixed areas that must not be monitored. Then there’s dynamic masking, which uses an Edge-based privacy shield application on visual cameras, allowing users to see movement or activity while protecting real-time privacy.

Another type of AI-based security solution is anomaly detection. AI is used to recognize behavioral patterns and to identify behaviors that lie outside of a learned norm. This is generally beneficial in combating users who may be accessing data for malicious reasons or who are in areas where they should not be. For example, it might be worth investigating a user who is constantly attempting to enter a secure area that they are not authorized to use.

Predictive analytics are another aspect of AI-based security solutions. By recognizing patterns, AI models can predict events that may pose a security risk. In the case of financial fraud, patterns of money laundering or other types of fraud can be analyzed using previous examples. Subsequently, the same patterns can be identified early on, potentially preventing people from becoming victims of scams or scams.

AI analyzes on the edge

With surveillance systems using edge analytics, it is possible for cameras to detect that something or someone is moving in a particular scene. This footage could then be analyzed by a human actor to understand exactly what the entity is and if it poses a security risk. However, incorporating AI analysis at the edge and training models to detect and classify different entities within a monitored area can yield incredible security benefits.

Performing analysis on an on-premises server has given us performance advantages. Now the powerful on-board processing offers new solution advantages at the edge. Edge analytics are video analytics that process and analyze video data right on the camera, close to where it was shot, not on a server or in the cloud. This type of built-in deep learning capabilities allows solution providers to offer unique capabilities such as: B. the development of AI-based third-party applications that can solve problems. Some of these apps can track people and send alerts when someone stays in a place for more than a certain amount of time, and don’t require people to be active or moving to be detected. And all this data runs entirely on the camera, without an additional server. It uses real machine learning to identify people and the length of time they were there. Basically, it tracks and sends alerts. This shows that server-based solutions are not the only option when AI or other powerful analytics are part of a security solution. Surveillance cameras equipped with DLPU (Deep Learning Processing Unit) chips, which have analytics installed directly on the camera, are more widely considered today due to their simplicity, scalability, flexibility and reduced cost.

READ :  Has the definition of "cloud" contributed to the challenges?

By integrating AI analytics at the edge, the rate of false positives can be significantly reduced, reducing the need for human intervention and more efficiently directing those resources to situations that require timely and appropriate responses. For example, surveillance cameras on highways, supported by AI analyzes at the edge, could clearly identify objects or accidents and warn drivers. These cameras could differentiate between vehicles and people, accurately alerting both drivers and emergency services to emerging situations in real time.

I’m looking forward to

As AI becomes more prevalent and evolving, so do cybersecurity threats. AI-based security solutions will eventually become a necessity in the modern world as they serve to protect individuals and organizations from threats. More research and development is needed to ensure that AI-based security solutions are public-focused and can be used responsibly.

In terms of edge computing, scalability and accuracy will only increase in the coming years. We’ve already seen tremendous advances in performance and computing power, so it’s exciting to envision the advances in object detection and analysis that are sure to come soon.