Wonder how the investigators in the Boston bombing are reviewing countless store, building, police and citizen provided videos?
Likely using technology and process the LEVA Forensic Video Analysis Response Team used to process over 5000 hours of video evidence related to the June 15, 2011 Stanley Cup Riot in Vancouver.
The review and tagging is a painstaking process. 52 analysts then spent 14 days processing that footage, identifying 15,000 criminal acts perpetrated by 300 rioters.
52 analysts @ 8 hrs/shift x 5 shifts/wk x 3 weeks = 52 analysts working 120 hours = 6,240 analyst hours processing 5,000 video hours = 1.25 analyst hour per 1 hour video review, suggesting all the value-added activities of review - ID, tagging, crime assessment, logging, etc -- takes 15 minutes per video hour. So the advantage of LEVA comes in that 15 minute chunk. I guess it could take longer WITHOUT LEVA, but one wouldn't exactly call this transformational. Or, am I missing something?
Posted by: Zach | April 18, 2013 at 07:02 AM
Zach, it is transformational in sense that it normalizes video from multiple sources/formats and allows for review virtually from multiple locations. Given the need for criminal evidence to be well documented it does that well. Clearly the analytical part could be expedited, and I am sure DARPA has analytical technology which is still classified. Crowdsourcing and Amazon Mechanical Turk could expand the review pool by a huge factor, but given sensitive nature of investigation and chance of bias, not sure it can be used easily.
Posted by: vinnie mirchandani | April 18, 2013 at 07:43 AM
Zach, at the RNC at Tampa, tools from AISight looked for certain patterns in video frames. Sure some of that is being used in Boston. see
http://florence20.typepad.com/renaissance/2012/08/the-department-of-precrime-at-the-republican-convention.html
Posted by: vinnie mirchandani | April 18, 2013 at 08:11 AM
Finally, an analyst or drone operator can access 65 screens simultaneously using tech like Argus so that should make pattern recognition that much quicker and better
http://florence20.typepad.com/renaissance/2013/01/the-argus-is-super-camera.html
Posted by: vinnie mirchandani | April 18, 2013 at 08:52 AM