Securicorp launches retrospective video data mining service

 

Video data mining is mainly instigated by law enforcement when they need to find footage of how and where an individual was and with whom in the lead up to an event.

A good example is when the 7/7 London bombings occurred. The individuals were quickly identified, but getting up-to-date video of the terrorists moving in and around London took about four weeks to put together. When tracking a suspect who has not been caught, this is far too much time when trying to get the public’s help. There are many reasons for this. First, there is the cameras’ management.

Identifying the cameras and then the location of where the associated recordings are can take time. Once that video is located, it is then removed to a central place for viewing. This viewing is a manual effort of watching the hours of footage, even the pitch dark night time video, by teams of people. Then, any video that is of interest requires further investigation or is important for evidence needs to be recorded onto another machine for viewing. The quality of the image is another consideration.

With video analytics software you can simply run the video footage through each channel with set parameters and use analytics based around what needs to be found. This parameter can be an individual crossing into a region, facial recognition, a color (of a car or a jacket, for instance). Once this is set, then as the video is run through the software, files of these parameters can be created for further analysis. One person can run multiple operations.

Securicorp would use its current Bikal Eyesoft and Bikal Biosurveillance and allow the rental of the appropriate servers as a service.