Inside the Analyst’s Mind

Is it a cave, or a weapons cache? Is it an equipment warehouse or is it an IED factory? These are the questions geospatial imagery analysts ask themselves in the course of their highly skilled work. Analysts study video and satellite images covering hundreds of square kilometers on a quest to identify objects, vehicles, military assets or individuals, all in an effort to provide our military and national security leaders with complete situational awareness. It is challenging and time-intensive work, which often amounts to searching for needles in a haystack. Fortunately, there is a robust, state-of-the-art geospatial-intelligence software tool - SOCET GXP - to help analysts process imagery more efficiently, and accurately. Interestingly, experts believe the next geospatial technology breakthrough already lies inside the minds of geospatial analysts. That’s why Columbia University Professor of Biomedical Engineering and Radiology, Dr. Paul Sajda, is examining how to maximize the capacity of the human brain to enhance analyst productivity.

Think of it this way, through rigorous training athletes can rely on muscle memory to elevate their performance. Dr. Sajda believes that this phenomenon can be applied to imagery analysts who rely on neural markers to “label” their work. These markers help the analyst prioritize what may be worth looking at, and what details may be worth exploring more closely.

“We want to help analysts identify and prioritize data that is relevant to their search,” said Sajda. “We can do that by harnessing the rapid recognition that humans are good at, without disrupting normal workflow. It becomes an extra output channel to record and annotate data.”

Over the past eight years, Sajda, and researchers at the neurotechnology and research firm, NeuroMatters, have been working closely with the Intelligence Community to study the human visual system’s ability to rapidly review and identify objects. Normally, imagery analysts review large satellite images from the top left corner, down over several minutes. But, during his research, Sajda broke the same large image into dozens of smaller sub-images. Then, he hooked the analysts up to a brain wave monitor, and asked them to review the series of images in succession. Sajda found that by studying the brainwave data, analysts could accurately process dozens of sub-images in a matter of seconds.

“If you were on a battlefield or an emergency room, you have to do a triage to effectively identify  threats, and determine how to prioritize them,” said Sajda. “From just a quick glance, imagery analysts are able to subconsciously identify objects of interest, which enables them to focus their time and attention on key areas of interest.”

Sajda often relies on the geospatial-intelligence software package, SOCET GXP, to produce the imagery for his research. Sajda says he trusts SOCET GXP because it is easy to use, and it is frequently used by many of his government clients. When imagery produced by SOCET GXP was used by experienced analysts in conjunction with brain scanning, Sajda says the efficiency of the average imagery analyst increased by 300%.

“We’re really trying to consider the big data problem,” said Sajda. “The number of tools and intelligence sources is increasing faster than the number of eyeballs reviewing them, so we are trying to come up with ways to tap into the experience of trained analysts to enhance imagery analysis.”

Dr. Sajda’s research was recently featured in a documentary produced by the BBC.