The Video Analytics Faithful
Three experts provide a partnership view of the analytics market and explain how companies are combining core competencies to harness the full power of video surveillance. Learn how scalable solutions provide capabilities beyond a limited set of predefined alarm situations, and where current and future growth opportunities exist.
I had an interesting conversation with SSI Editor-in-Chief Scott Goldfine at the end of last year regarding this magazine’s editorial agenda and the hot topics expected to emerge in 2010. We discussed the rise of megapixel and HDTV, the continued transition from analog to IP, and the inevitable growth of hosted video to bring the latest in security to the masses. We traded integrator and end-user stories and appeared on the same page regarding the major trends.
But, as we broached the topic of analytics, Goldfine asked with a wry tone, “Do you think this will finally be the year that the analytics market takes off?”
It was a candid question. In the mid-2000s video analytics created a stir throughout the industry. Many companies emerged and promised analytic capabilities that were, frankly, a little ahead of their time. These complex analytic functions — such as bag-left-behind and facial recognition — would require specialized teams, engineering and many years to develop; yet some companies were claiming these capabilities were right around the corner.
Now that some of the early hype surrounding video analytics has subsided, major industry players have developed a more realistic view of what is and isn’t possible with analytics. Better yet, as I explained to Goldfine, companies are successfully partnering together to collaborate on ideas and technologies so that the future of analytics might soon match its early hype.
Since much of this collaboration is done behind the scenes, I thought it would be enlightening to share an open discussion with two of Axis analytic partners. What follows is an in-depth conversation I recently had with Justin Schorn, vice president of product management at intelligent video management solutions provider Aimetis Corp., and Joachim Stark, global director of digital video surveillance at IBM. Both men kindly agreed to discuss the analytic market and how application development partners are building useful and cutting-edge functionality via an open platform.
How would you describe the state of the analytic market?
Joachim Stark: After quite a few years of research-level progress, there was a phase where clients questioned the validity of their return on investment [ROI] from video analytics. However, over the past 12 months this situation has started to improve. We now see projects being scoped at a more realistic level, with the development of improved technology and solution services. Clients now see the value coming from a flexible and highly scalable solution that provides video intelligence on a more generic level, beyond the capabilities to identify only a limited set of predefined alarm situations.
Justin Schorn: I agree that end users are starting to see the clearcut benefits, but believe that the analytics market is still in its infancy. Unfortunately, analytic expectations were not properly managed early on, breeding discontent from the early adopters. Nonetheless, the industry is slowly restoring confidence with many successful reference projects. Armed with more in-camera processing power and past experiences, analytics technology is on the right track to eventually become a profitable commodity.
We can all agree that expecta-tions were poorly managed early on. But as partners, how can we better commu-nicate to the integrator and end user that we’re improving the marriage between an open camera application platform and advanced analytics to offer func-tional tools for security and busi-ness?
Schorn: The analytics solution space isn’t always well understood by the integrator or end user. Understanding where we can be successful and where not is part of the problem, but this can be solved through better education and train-ing. Many integrators are not properly trained resulting in improperly configured analytics deployments. As a result, a lot of potential value is lost due to high false alarms or inaccurate business intelligence reporting. Analytics vendors, camera platform vendors, and integrators alike must invest in the necessary education and training to bridge this knowledge gap.
Stark: Aside from training, there is a strong need to improve and partially automate the installation, inte-gration, tuning and operational/post-installation support of the video analytics. Designs to improve the installation and management of these systems are becom-ing a focus for technology and services development partners. This ultimately leads to a better product and service infra-structure, which can be integrated into solutions in a faster, more predictable and reliable manner, delivering increased ROI faster.
Do you agree that by moving some of the analytic capabilities to the edge of the network — i.e. on the camera — that ROI and productivity will increase? From the developer’s point of view, are there keys to delivering effective analytics ‘at the edge’?
Stark: Basic analytics like camera manipulation and system outages can easily be handled on the edge device — on the IP camera itself. As edge devices continue to provide increasing computing power, some more advanced algorithms can then run on the devices themselves, without the need to first send compressed video across the network to a video management solution [VMS], spend computing power to decode it again, and have vision computing algorithms run on a server. This saves both power and time. Where additional complex analytics algorithms are needed, centralized server-based execution can complement the edge-based analytics.
Additionally, when edge devices run analytics algorithms in a more generic manner and transmit the corresponding metadata describing more or less any object movement beyond predefined alerts, this can create a real-time, postevent decision support for security operators through an advanced back-end correlation system. So, the key to deliver effective analytics at the edge is having the capability to do generic video analysis and to provide the metadata to a backend system that can act as an ‘intelligence store’ for data correlation.
Schorn: The advantages to running analytics at the edge are many. Foremost is scalability. Users need only add another camera and don’t have to add more computer power to the server. Another is isolation. If one camera goes down it doesn’t crash the whole system. Finally, installation is easier. No extra wiring/hardware needed for analytics installation since it resides on the camera. However, the analytics need to fit within the computing resources of the device. Even if it does, centralization of metadata is likely needed and a strong VMS product at the backend that can properly leverage the analytics metadata is a must.
If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our FREE digital newsletters!
Related Content
Security Is Our Business, Too
For professionals who recommend, buy and install all types of electronic security equipment, a free subscription to Security Sales & Integration is like having a consultant on call. You’ll find an ideal balance of technology and business coverage, with installation tips and techniques for products and updates on how to add sales to your bottom line.
A free subscription to the #1 resource for the residential and commercial security industry will prove to be invaluable. Subscribe today!
