The Transportation Security Administration’s (TSA) issuance of a Request for Information (RFI) to take a look at a wide range of intelligent video capabilities is welcome news to firms in the area who believe the recent notice from the agency is an acknowledgement that the technology has arrived.
The RFI is “unique in the breadth” of what TSA wants to look at, Dvir Doron, vice president of marketing with ioimage, an Israel-based supplier of intelligent video appliances, tells TR2. He says the fact that TSA is showing a great deal of interest in the technology “will rock the video analytics world” because it is a validation of the technology from an important homeland security agency.
TSA realizes that “the technology is coming into the mainstream” and with the responses “they’ll find a whole range of capabilities are available, from laboratory development to wide scale deployments,” Craig Chambers, president and CEO of Cernium Corp., tells TR2. Cernium has two primary video analytics products, ExitSentry, which is used in over 40 airports to automatically detect and alert for people entering an airport exit lane from the wrong direction, and Perceptrak, which validates the presence of objects, vehicles or people and looks for patterns that could represent a threat.
In addition to more of the video analytics being accepted for security uses, the price points are becoming more attractive, Ty Sellers, a regional sales manager for Siemens Building Technologies [SI], tells TR2. The technology is allowing for a more efficient use of capital resources, he says.
TSA says in its RFI that it is focused on finding solutions for aviation security but is interested in eventually having the technology applied to other transportation modes.
While TSA is hoping it can use the technology to add another layer of security in the transportation environment, it also says it wants to see if it can “identify and attempt to resolve the real world problems associated with intelligent video-based systems” and “create an additional path going forward to provide for the installation of intelligent video-based systems within our transportation systems.”
TSA has taken the opportunity to look at video analytics previously, most recently last fall when it issued an RFI for technologies and operational capabilities to detect people passing through the exit lanes near airport security checkpoints (TR2, Oct. 17, 2007). However, that RFI was open ended in terms of the potential solutions.
In its new RFI, TSA is taking a big bite at what solutions may be available. For example, the agency would like responses on technology that can do macro-behavior detecting, such as large crowd level anomalies, walking in the wrong direction, loitering, object classification, and others. TSA would also like to know if the technology exists to automatically track individuals across multiple cameras.
At the micro-level of behavior detection, TSA is seeking responses related to automated micro-facial expression recognition that might indicate stress, nervous-related actions such as sweating, or other anomalies.
Cernium’s Chambers says from his experience in the video analytics industry the micro-level behavior detection technology is still in the laboratory stages. A key challenge is the fact that the level of video quality to be able to detect this behavior is poor, he says.
Paul Ekman, a pioneer in the field of research into micro-facial expressions, wants the opportunity to evaluate the video analytics technology for detecting micro-facial expressions. Micro-facial expressions typically last a fraction of a second and occur when someone is trying to hide an expression. Ekman’s firm, The Ekman Group, has trained TSA inspectors to observe micro-facial expressions (TR2, Aug. 8, 2007).
Ekman cautions that “detecting micro expressions automatically may be hard to do, harder than detecting macro expressions, and even if accomplished won’t identify malfeasants.” Moreover, he tells TR2, “it is only one clue and alone not always useful unless combined with gesture, posture, gaze direction and other observables which as yet can’t be detected automatically.”
The agency is also interested in suspicious item detection such as abandoned objects or known threat objects. Additionally, TSA wants to see if the technology is available to bring the various inputs from intelligent video systems together to create a common operating picture, including the use of wireless handheld devices for video and data for remote operators.
CCTV systems have been in use for security in the transportation sector since the 1970s, allowing security personnel in central locations to monitor activity in select areas remotely. But as the technology is relied on increasingly, it also requires more security personnel to monitor the imagery. The potential of intelligent video analytics is that sophisticated software algorithms can process the imagery and automatically feed alerts to security personnel when there is some sort of anomaly or security breach.
Providing automatic alerts should just be the beginning, says Chambers. To become a real force multiplier that information needs to be aggregated and presented in a larger context for scenario management, he says. Ultimately, this is what he hopes TSA’s goal is as it begins to examine the potential of intelligent video.
What TSA ultimately decides to do after taking its first significant look at intelligent video isn’t clear. Greg Hull, the director of security for the American Public Transportation Association, says that for the non-aviation sectors TSA won’t buy any of these systems. As it stands, he says, more restrictions are being put on grant requests that transit agencies may submit and they face the prospect of even higher cost sharing requirements. Still, he adds, some transit agencies have began using smart software for their CCTV systems, which need to find ways to get more use out of a limited number of people. [Sol. No. HSTS04-08-ICT7017. Respond by March 14. Contact: Connie Thornton, contracting officer, [email protected].]