Researchers and organizations alike are capturing data through sensors and cameras that leave them awash with information, though they are often only interested in small details. In the quest to uncover the Higgs boson, researchers produced 25 petabytes of collision data annually, and only a tiny fraction of that contained useful anomalies. NASA alone produces an average of 1.73 gigabytes of data each second. (If you’re curious about just how much data could be created in a year, take a look at “What is Big Data and why does it matter?”by Tom Soderstrom, chief technology officer at the NASA Jet Propulsion Laboratory.)
When it comes to analyzing high-resolution images, the human eye is still an invaluable tool. However, when humans are involved in analyzing extreme scale data, something very limited – human attention– is allocated throughout something very large, which ultimately becomes a zero sum game. “If you focus more attention on one part of the data, you are ultimately taking attention away from another part of the data,” explains Amitabh Varshney, director of the Institute for Advanced Computer Studies at the University of Maryland, US.
Working with perceptual psychologists, Varshney’s lab carries out experiments using eye trackers aimed at determining exactly what it is that humans see (as opposed to what they don’t see). “The human visual system looks at contrast in various dimensions: luminous contrast– brightness vs. less bright, textural detail, orientation, changes of curvature and motion, and it does this all in a particular order. If we can identify the most salient areas that human eyes are drawn to, we can create a set of visual tools and image processing algorithms to guide human attention.”
Our eyes not only trigger visualization and bring visual input to our senses; they activate memory as well. “Visualization is a key component in helping augment memory. It allows us to transfer processes that are fundamentally cognitive – thinking and decision making tasks – to much more perceptual processes,” explains John Stasko, associate chair of the School of Interactive Computing at the Georgia Institute of Technology, US. “Visualization helps to engage our pattern matching system, which we’re really good at; machines are good too – but humans still have an edge when it comes to this.”
Stasko is also the director of the school’s Information Interfaces Research Group, which creates tools that combine computational analyses and interactive visualization. Jigsaw, one such tool, is a visual analytics system designed for exploring and understanding document collections, essentially creating visual threads that weave through documents in various ways.
Stasko actively seeks feedback from various organizations that use Jigsaw, which is freely available online. Organizations and agencies that have benefited from the tool include law enforcement, federal agencies, fraud investigations, investigative reporting, and consumer reporting.
Just this month a 320-gigapixel image of London, taken from the BT Tower, set a new record for the world’s largest panoramic photo. The photo is made up of more than 48,000 individual frames that have been stitched together by a supercomputer. The individual images took three days to shoot and compiling them into a single panorama took three months.
All told, the flow of high-definition data is unlikely to slow any time soon. Last summer researchers at Duke University in Durham, North Carolina, US, put the finishing touches on a 50 gigapixel camera with a massively parallel array of electronic elements. They expect a handheld version to be available to the general public within five years.