Algorithm sniffs out graffiti artists

Gangs tend to mark their territory with graffiti, as well as leave signs to communicate with those in the loop about drug deals, and the police are left scratching their heads as they work on deciphering those “messages”. Well, the long arm of the law has just received a brand new ally in the form of an automated graffiti-identification system that might actually be able to figure out who the perpetrators are, making their job a whole lot easier.

At the moment, the men in blue in the US manually label images of known gang graffiti using relevant keywords, including the gang name. To go through the collected database of such images and to match them with new graffiti is a tedious affair, and Anil Jain, a computer scientist at Michigan State University in East Lansing, decided to help the police by working on a system which is smart enough to make comparisons photos of graffiti, picking out similarities and identifying the gang that each piece belongs to. Sometimes, the system during its trial managed to even identify the individual who made the graffiti – now how about that? Art has been broken down to bits and bytes now, although it is still not foolproof, but it sure makes life easier for the police.

Edwin Kee
Categories: Computers
Tags: police