The US Defense Advanced Research Projects Agency (DARPA) Information Innovation Office (I2O) (Arlington, VA, USA) has released details of software it would like to see developed to analyze information from photos and videos held by the US Department of Defense (DoD).
Insurgents and terrorist groups frequently use video, still, and cell phone cameras to document their training and operations and occasionally post this content to web sites. But the volume of this visual media is growing rapidly and is outpacing the ability of individuals in anti-terrorism agencies to review -- let alone analyze -- the contents of every image to assess threats, obtain evidence, and find the location where such activity is taking place.
To help with that task, in its request for proposals (RFP) DARPA envisages that the Visual Media Reasoning (VMR) program would be able to automatically extract relevant information so that human analysts could then be alerted to any scenes that warrant their attention.
The software application would enable users to submit photos or video frames of interest and indicate what they would like to know about the contents. The VMR system would then interpret the image, after which the user would receive a set of possible responses, with the system’s assumptions and/or likelihood of certainty made clear.
DARPA will hold an Industry Day briefing on the VMR program on August 31 in Arlington, VA.
-- Posted by Vision Systems Design