Automated Image Analysis Video Will Be the Next Big Breakthrough

Remember the character Chance Gardener in Being There? "I like to watch TV" was his famous line. The problem with video is that most companies produce a lot of unwatchable movies, especially those taken from security camera systems. And there are probably around who want to see them, either.

The problem is that trying to find one or two enjoyable events in all this movie is not easy. As an example, take the story in the New York Times today, which reminds us of the military to be buried under a huge pile of movies from its Predator drones that are flying in Pakistan and Afghanistan. The situation is only likely to worsen as the newer models of the drones will be sending streams from dozens of cameras in a few years.

The story has already been said about how video feed of the drones' are available to anyone with a few minor software skills - Effective control of the channels are encrypted, but the video streams are not. This is because most legit people need to see what they are transmitting and the army was not able to implement all packets encrypted visualization of these flows.

In the Times article, a group of soldiers based in Hampton, Virginia to sit in front of the screens and see in real time, and then make them available to the screengrabs right through computer chat room. I hope for the good of all concerned that these chats are encrypted, but the article did not say.

So how do we implement our automated Chance Gardner? There are a couple of technologies that can help here, but it is not easy or cheap to implement. One is the telestrator, the device made popular by John Madden and Monday Night Football in which one commentator draws on the screen and you have seen colorful doodles to show what is happening. Madden and those used by professionals are very expensive, but there are dozens of products available for the PC market, with some freeware products as Producer VideoMage.

The telestrators are nice, but again, someone must be watching the video and make the electronic squiggles. Need more than in fast-forward button to do this - ideally, you want some kind of automated system capable of identifying feasible moments on video. This is what the next class of products called Intelligent Video Analysis. They look at the flow computer and highlight the particular activities that a human operator can go back and review later.

This is what stoplift.com company is doing with its analysis systems retail checkout. Typically, a shop to install cameras on each checkout lane and record what you are doing as a corrector elements pass through the point of sale scanning devices. There are all sorts of scams that can be used as "sweethearting" (an accomplice is giving free items that are not scanned) and watch as you scan a barcode when you are just passing the ' element around and above the bagger. So, what is needed is a system that ties into your store and can tell you when these items are not called to the cashier. I got to see a demo last week and thought this was so cool. The company claims its software can have a period of six months ROI and significantly reduce the cost of goods stolen. And the good news is that nobody has to watch all the security tapes to see just love those moments....