The Agency for Advanced Research in the Sphere of Intelligence (IARPA) allowed leakage of information about the project “Finder”. A special search engine, whose task is to analyze an infinite number of user photos and determine the geographical coordinates of the place in which they are made. Of course, with the good purpose of quickly discovering the enemies of the state who adore the Selfie.
Analysis of the geotagging of user publications as a whole was adopted by special services a long time ago, but technically literate people have learned to disconnect them in their gadget settings. Therefore, the new concept is to calculate the location of the author of the snapshot by indirect data, and then assign it the same geocode, but only in the internal spy system, for official use.
The predecessor Finder was a neural network from Google, known as “PlaNet”. With its help, 126 million photos were processed in open terrain with and without geotagging to teach the neural network to recognize the landscape in the background in the frame. The results were good, but scanty – if in the same East Gut in Syria little photographed as a whole, then there was nothing to work on neural networks either.
Therefore, the new goal of the Finder project is to learn how to use data from any sources in order to determine as precisely as possible the location on the planet Earth where a particular image was taken. Among the contractors are Applied Research Associates, BAE Systems, Leidos, the search engine will be compatible with the tools Aladdin Video and Deep Intermodal Video Analytics. The first analyzes the metadata of video clips from social networks, the second detects the signs of “dangerous actions” of people on the video and can use multiple cameras at the same time.