Friday, December 09, 2011

#ALGORITHMS: "Video Analytics Boosts Business Intelligence"

The millions of video surveillance cameras worldwide have opened a nearly billion-dollar market opportunity for smart software that can detect motion, recognize people, identify objects and track vehicles.

Analytics running on the video feeds from the millions of surveillance cameras worldwide are recognizing objects and their behaviors to provide situational awareness of remote locations. As these algorithms get smarter, a new market is forming that incorporates video analytics into business intelligence tools. This market has the potential to grow to nearly $1 billion by 2016, according to ABI Research.
Business intelligence tools incorporating video analytics will become a $900 million market by 2016, according to ABI Research.
Video analytics gets its smarts from emulating the functions of a biological visual cortex, allowing it to extract features from images with recognition algorithms. That information is then used to deduce the perceived behaviors in a manner similar to how a human would. Today security is still the seminal application, but business apps are growing fast in fields as diverse as entertainment, health care, retailing, automotive, transportation, home automation and safety.
Most video analytics today run on servers, but accelerated PCs are also getting into the act, moving the analytic engines closer to the "edge" of the network where the video cameras themselves are located. Starting out with simple security algorithms, such as detecting motion inside a retail business at night when no one should be there, video analytics are getting smarter. For example, today they might be used by an operator to pinpoint objects that are moving, outline them with bright colored outlines and track them as they move.
For retail, just counting the number of people frequenting different sections of a store can be useful, but more sophisticated algorithms are already available. For instance, digital signs are starting to use cameras to perceive their viewers, and then use analytics to change the displayed content to match the age and gender of the person currently looking at the sign. For sports, video analytics are being used to follow the ball, highlighting the player currently in possession as he or she moves around the field. And for mobile applications, smart video egomotion estimation techniques are able to identify the route being driven from vehicle-mounted cameras, creating a kind of visual odometer that does not require GPS.
ABI Research recently surveyed smart video analytics in its report Intelligent Video Analytics, where it concludes that business intelligence usage of video analytics is shifting to PC-based edge-located devices, an effort being led by Intel's digital signage group.
Further Reading