SSI spoke with Western Digital Director of Segment Marketing Brian Mallari recently to highlight some trends in smart video deployments that 173 security integrators weighed in on.
What are some of the benefits of pairing AI with camera systems and DVRs, and how’s that changing what security providers have done for so long, which is recording of events and reviewing the video later?
Brian Mallari: There are generally two ways in which AI can help in a video surveillance solution. First, AI enhances the effectiveness of a security solution. The idea is that cameras can now see things as good as, or even better than humans have been able to in the past. Some functions such as watching the perimeter for intruders or distinguishing between a human, a raccoon, or a dog or scanning for violent behavior — those kinds of things can be entrusted to smart cameras.
This allows more effective use of security personnel, who would no longer need to stare at cameras all day. The cameras can alert the security personnel in the event something happens, so it makes for more efficient use of one’s time. That’s one way that AI helps with traditional security.
Another way is in nonsurveillance-related use cases, which are starting to emerge. For example, AI brings a greater ability to extract insights from video, and that allows an end customer to drive improved business or more efficient operations. Case in point: analysis of hours of video from a retail store can determine the paths most frequently used through the store, and these resultant “heat maps” can help the retail business place new or most profitable products more effectively. Whether it’s in a corporate setting, manufacturing, retail, city management, healthcare and so on, there are a lot of tremendous opportunities for AI-based video, outside of the typical security realm.
You actually just commissioned a recent survey of security integrators and covered a few things about how businesses want to use AI. Can you share what some of those are, and which are growing faster than others?
Mallari: Yes, Western Digital conducted our second annual smart video survey with help from Security Sales & Integration. It surveyed more than 170 respondents sharing use cases for AI-based video solutions, based on what is resonating with their customers.
I think the biggest benefit of AI is that it ultimately drives better business decisions. From the survey, we found that integrators and architects of security solutions today are seeing the most use of AI-based video solutions in the areas that you might expect — the top category in the survey being businesses/technology/enterprise campuses. The next most AI use cases were deployed by banks and financial institutions, followed by casinos/gambling facilities, and then, government facilities. Whether AI is used on assembly lines, or for building access control, or for onsite, or perimeter security, these verticals have been the early adopters of AI in video technology.
Beyond those verticals, the survey also found colleges, universities, smart city, retail and healthcare use cases were rising. When we asked the integrators how things have changed in the most recent months, the results changed slightly. The verticals that I mentioned remained generally consistent, with the exception of smart city deployments in AI taking a spot higher in the rankings, moving to sixth place. It’s reflective of the growth and interest in smart cities.
Did you find any surprises in the survey?
Mallari: I wouldn’t call it a surprise, but the survey found that there’s a need for education on best practices for deploying AI-based systems. Again, this survey was directed to integrators, and so the education that people were talking about had to do with basic things such as how to design a system using AI, and providing examples they could work from. There was also lack of understanding on who makes the decision for AI use, and how the deployments are funded. This was particularly true for new and advanced uses of AI for which there is strong interest. It’s a gap that we identified, and a gap that we can help close as the trusted advisor to our customers.
Storage is a major driver for these new use cases in smart video. Western Digital works with top integrators who use our purpose-built storage for surveillance and video – the WD Purple products – to build more advanced solutions. We’re working to turn those solutions into case studies to help others, further addressing the gap between deployment interest and the need for education. With our partners, we are helping integrators design systems that make sense for their customers’ requirements. We have the benefit of many partners that we engage with to deliver these advanced solutions.
When it comes to storage, more storage is needed to handle AI workloads, but there seems to be a sweet spot of sorts in today’s deployments, according to the survey. What capacity and storage hard drives are most used, and will that grow and why?
Mallari: The survey showed us that nearly eight out of 10 integrators say that new systems are deployed with larger capacities already. Most said they deployed systems using 4-8 terabytes of storage (42.2% of responses); and the next highest number deployed systems using 8 terabytes of storage and above (34.9% of responses). As AI continues to gain adoption and use cases evolve, we do expect the average amount of storage used in each deployment will grow. The reason for that is AI effectiveness improves with higher video resolution.
Higher resolution video drives more effective extraction of insights from video. For example, a license plate is clearer using 4K video versus 1080p, so the chances of the AI algorithm recognizing a license plate increases with higher resolution. And so, we see the market moving from 1080p cameras to 2K cameras and then to 4K over time. Higher resolution video drives higher bit rates, which in turn drives up the need for storage.
Finally, it isn’t just about video resolution. As AI grows, there’s a lot of AI metadata that’s created. I am talking about metadata that is not just in the video stream, but also metadata related to AI. New cameras have more than a main stream. They also have what are called auxiliary streams and picture streams. Picture streams are actually pieces of the video – meaning a frame of that video is taken about three times a second and stored, and all those pictures are used in some AI functions.
This data is useful for pattern matching, for example. So, you can see that the AI in a camera is doing more than just storing videos. It’s actually storing all of the other metadata and files needed to make a more effective AI solution — one that brings insights too.
What can integrators and those deploying these smart video solutions do today to get their customers educated and ready to get more benefits and visual data?
Mallari: There is a significant value that businesses could be capturing from deploying AI in their facilities, which could include improving response times to critical events, enabling more sophisticated analysis of visual data or analyzing patterns to improve business intelligence. We believe that educating integrators and customers about these benefits is an essential starting point.
AI sounds sophisticated and some people believe it’s something only large corporations or governments building smart cities can benefit from. However, we believe that most businesses stand to benefit from incorporating AI into their day-to-day operations. As businesses get up to speed on AI, we are working to assist our own customers. With security and systems integrators, it’s both about knowing the new use cases and getting the knowledge and training they need, which this survey is helping to made clear.
What thoughts can we leave our readers with as they begin to learn more about how to add AI and machine learning algorithms to their deployments?
Mallari: In close, I can say that while some of those polled felt that a classic surveillance deployment is good enough for their needs today, they see the value of AI and they simply need guidance to get started. We believe that with improved capabilities, such as higher resolutions and improving equipment costs, AI systems are actually able to serve many more businesses and organizations than was possible just a few years ago. Even the advanced NVRs today are starting to ship with built-in AI capability.
We at Western Digital want to engage with more integrators and end users to highlight how AI can help deliver more business value. We also want to help move away from the idea that AI is controversial, but instead, really focus on the fact that it provides an exciting way to improve their operations. Better use of visual data creates new business opportunities, which can deliver more value for the customer.
To view the full findings from the survey, click here.
The post Western Digital: Businesses Seek Advantages Through AI-Based Video Solutions appeared first on Security Sales & Integration.