Software predicts people about to commit crimes

US scientists are studying a software that can identify the actions of people on surveillance cameras and use algorithms to predict those who are going to break the law.

In Minority Report, an American film released in 2002, police forces in a US city know ahead of the killers by predicting the future of three mutant mutant mutants.

Carnegie Mellon University software can analyze basic behaviors
of people to predict objects that are about to commit illegal acts.

It was just a fictional film, but it was possible that the prediction of the offender would come true through the efforts of Carnegie Mellon University scientists in Pittsburgh, Pennsylvania, USA. Recently they have published a report to explain the working principle of the system to predict illegal behavior. According to the report, a software will process image data from surveillance cameras to detect suspicious behavior of humans, Physorg reported.

The software can identify actions such as running, stepping, calf, picking, pulling, following, chasing. It then takes action into the surroundings and surroundings to predict the likelihood of a crime. If a suspicious object is detected, the software will send a warning message to the police. For example, a warning message will be sent if a person carries a heavy bag and leaves it outside a restaurant.

With Carnegie Mellon University software, computers can replace people in processing data from surveillance cameras. The cost for computers is much lower than that of people. In addition, computers work continuously, do not fall asleep, disperse thoughts like people so do not miss any suspicious cases.

The system predicts criminal acts may appear in public places in the future, such as airports, bus stations, and stations. The team's goal is to help surveillance cameras prevent illegal behavior, not just broadcast warning messages.