July 31, 2013: Intelligence analysts finally have software fast enough to rapidly scan video images for specific data. This is done by taking high-end video cards and using their image processing capability and speeding up analysis by using similar cards tweaked to operate like a super-computer. In this way analysts can study video data up to 75 times faster than with current equipment. While this sort of capability is critical for military operations, it is also in demand by police and for a wide array of scientific research.

While the military has long been a major user of supercomputers (the fastest computers on the planet), they were too expensive for widespread use until graphic card maker Nvidia developed and sold graphic cards modified to be cheap and powerful supercomputers. Seven years ago the Tesla supercomputer add-on for PCs appeared on the market. This was basically an Nvidia graphics board tweaked to act like a supercomputer, rather than a device that put 3-D, photo-realistic game graphics on your computer screen. The latest version of this system will give you over five teraflops of computing power for under $7,000 (on one Quadro 6000 card, basically a tweaked graphics card). Using cards like this, some of the fastest and cheapest supercomputers in the world are built. Of the current list of 500 fastest supercomputers in the world, one powered by Tesla cards holds position 52. The new GeoInt Accelerator version, combined with Tesla and special software, allows analysts to scrutinize more data faster, meaning the likelihood of finding something useful went way up.

Supercomputers were first developed, as were the first computers, for military applications. These ultra-powerful computers are used for code breaking and to help design weapons (including nukes) and equipment (especially electronics). The military also needs a large amount of computing power for data mining (pulling useful information, about the enemy, from ever larger masses of text or visual information).

Because there's never enough money to buy all the supercomputers (which are super expensive) needed, military researchers have come up with ways to do it cheaper. Back in the 1990s it was military researchers who figured out how to use GPUs (Graphic Processing Units, from high end graphic cards) for non-graphic computing. GPUs do something similar to what supercomputers do (lots of math calculations of a fairly simple type) and eventually the manufacturers of GPUs realized that there was a commercial (not just military) demand for GPUs serving as supercomputers. Currently over ten percent of Nvidias video cards are Teslas and that is expected to quickly go north of 15 percent with the introduction of the new GeoInt Accelerator version.

This is what the military needs to make video analysis software practical. The U.S. Department of Defense has been developing two new software systems that can take over the tedious job of watching video feeds from UAVs or security cameras. One is VIRAT (Video and Image Retrieval and Analysis Tool). The other is PERSEAS (Persistent Stare Exploitation and Analysis System). VIRAT is for video stakeouts of small areas (a building or single doorway or window), while PERSEAS collects activities over a wider area and uses statistical analysis, and databases of enemy activity, to look for useful patterns. This stuff worked but until something like GeoInt Accelerator came along, was often too slow to be effective.

This is the continuation of a long term trend. For over a decade the U.S. has been using software to help scour satellite and aerial recon pictures for useful information. There were simply not enough trained photo analysts to examine the growing number of photos generated in the course of intelligence work. The boredom of watching video for hours is increasingly being alleviated by the use of pattern matching software that can detect movement that is in need of human attention. Research has shown that people staring at live video feeds start losing their ability to concentrate on the images after about twenty minutes. This problem has been known for some time, and the military (not to mention civilian security firms) have been seeking a technological solution. It's actually not as bad with UAVs because the picture constantly changes but cameras that are staring at the same spot can wear operators out very quickly.

The basic tech solution is pattern analysis. Since the most common video is digital, it's possible to translate the video into numbers, and then analyze those numbers. Government security organizations have been doing this for some time but after the fact. It's one thing to have a bunch of computers analyze satellite photos for a week, to see if there was anything useful there. It's quite another matter to do it in real time. But computers have gotten faster, cheaper, and smaller in the last few years, and programmers have kept coming up with more efficient algorithms for analyzing the digital images. Commercial firms already have software on the market that will analyze, in real time, video and alert a human operator if someone, or something (you are looking for), appears to be there.

The real time analysis software is rapidly evolving. You don't hear much about it because if the enemy knows the details of how it works, they can develop moves that will deceive it (or, to be more accurate, make the pattern analysis less accurate). Already this software is being used as an adjunct to human observers and gradually taking over. There will always be a human in the loop, to confirm what the software believes it has found. The new systems possible because of the GeoInt Accelerator/Tesla combination allows older video to be scanned faster than real time, allowing a lot of valuable information to be extracted from video taken years ago.

The proliferation of video cameras on the battlefield (in UAVs, ground robots, and for base security) has provided a huge library of images that show bad guys doing what bad guys do. This can range from moving around carrying weapons, to using those weapons, to the particular driving patterns of people up to no good. This is a unique resource, and the U.S. Department of Defense has put together a library of these images. This is similar to older still pictures libraries, which were eventually used by pattern recognition software to let machines examine the millions of images digital photo satellites began producing decades ago. The basic problem was that there were soon too many pictures for human analysts to examine. Computers had to do much of the work or else most of the images would go unexamined. This technology was quickly adapted to the kind of combat encountered in Iraq and Afghanistan and terrorist operations in general.