You and another 59,000 people are watching a tense football match in a crowded, rowdy stadium. Your safety depends on a group of people responsible for ensuring your security. But what happens when things get out of control and security guards can't pinpoint the problem? Enter 'Smart Eyes', an EU-funded state-of-the-art camera system with the potential to boost security in public buildings and areas. This special surveillance system acts like a human eye by analysing the recorded data in real time and identifying out of control events and offering key solutions. The camera is an outcome of the EU-funded SEARISE ('Smart eyes: attending and recognising instances of salient events') project, which has received EUR 2.15 million under the 'Information and communication technologies' (ICT) Theme of the Seventh Framework Programme (FP7). According to the SEARISE partners, led by the Fraunhofer Institute for Applied Information Technology (FIT) in Germany, the automatic camera system can detect, track and categorise salient events and actions. The Smart Eye system is so similar to the human eye that it can distinguish objects when looking at a scene regardless of the activity around these objects. Video data is assessed in real time, and the camera 'points out salient features', the partners say. 'That is invaluable for video surveillance of public buildings or places,' explains Dr Martina Kolesnik from FIT. 'In certain circumstances the capabilities of a human observer are limited. Ask someone to keep any eye on a certain stand in a football stadium and they are bound to miss many details. That same person can only carefully monitor certain sections of the whole area and will quickly get tired. That's where the Smart Eyes clearly comes into its own.' The system hardware developed by the SEARISE partners encompasses a fixed surveillance camera capable of covering a specific area and two ultra-active stereo cameras. Just as human eyes fix on and track objects and points quickly and pointedly, the Smart Eyes system can as well. Launched in 2008, the three-year SEARISE project has built a software system with a hierarchical, modular structure (a computational model of visual processing in the brain) that automatically analyses the image sequence. Identifying each pixel movement, the software helps the system pinpoint particularly active areas in the scene. According to the team, the system learns motion patterns and stores them as typical models, and then uses the models to identify and classify events. A case in point is that the software can differentiate between active and passive spectators. While the program identifies image patterns like steps or empty seats, it also filters out various objects including flags. 'Our image analysis software is compatible with camera systems produced by all vendors,' Dr Kolesnik says. 'It can be installed easily. The user doesn't have to make any adjustments.' The Smart Eyes system will be on show for the public at the Security Essen 2010 exhibition from 5 to 8 October. The SEARISE partners are from Germany, France, Italy and the UK.
Germany, France, Italy