To mark their territory and warn off rivals, 21st-century gangsters still depend on the street language of graffiti. “Graffiti is a big part of how gangs tell their story and pick their turf,” says Steven Schafer, a detective in the criminal gang unit of the Indianapolis Metropolitan Police Department. A new software program called GARI (Gang Graffiti Automatic Recognition and Interpretation) is now helping Schafer and other investigators decipher the scrawlings, monitor gang activity, and fight crime.
GARI connects officers in the field with a searchable database of graffiti information and images snapped by cell phones and digital cameras. An officer can take a photo and submit it to an app, which tags it with location, date, and time. The software also scans the graffiti for distinguishing features, including color and shape. Officers can then enter queries into GARI to check for similar images logged within a certain area and derive local gang affiliation, territorial disputes, and even the identity of the members who left their mark.
Because GARI is so new, Schafer and his team must manually tag many of the submitted photos to build up the app’s information bank. The program is also still honing its ability to identify graffiti on a variety of materials, from wood to dirty cement. But even in these early stages, more than a dozen police agencies in Indiana have signed up with the program. “The real challenge is acquiring and processing images,” says Edward Delp, an electrical and computer engineer developing GARI with researchers at Purdue University. “It’s not like reading a sign on the street—every image is different.” The photos displayed here have been fed into GARI as part of ongoing police investigations in Indianapolis. The app itself is not available to the public—including members of gangs who could use it to avoid getting caught.