Facial Recognition Software Helps Bengaluru Police Identify 2.5 Lakh Criminals in 90 Days

Facial recognition software deployed by Bengaluru police has identified 2.5 lakh individuals with criminal backgrounds over the last 90 days, leading to the arrest of at least 10 people. This development comes under the Safe City Project, where advanced technology, including Facial Recognition System (FRS), is being used to boost city-wide security.

Safe City Project and FRS Deployment

The FRS was introduced on a pilot basis in Bengaluru as part of the Safe City Project, and it’s currently linked to AI-powered cameras set up across the city. These cameras send live feeds to the police command centre, where the FRS software matches faces captured on camera with the criminal database maintained by the Karnataka police.

Police Commissioner B Dayananda explained that two different facial recognition software are being tested for the system, and the most effective one will be adopted. “The ability of the system to accurately match a subject’s facial features with the criminal database needs to be stabilised,” he told Deccan Herald. The current accuracy rate is between 75% and 80%, with the goal being to achieve 98-99% through machine learning and data enhancement.

How the System Works

Around 7,500 AI-powered cameras have been installed throughout Bengaluru, of which 1,000 are linked to the FRS. These cameras track live feeds and notify the command centre when someone with a criminal background is detected. In some cases, like the apprehension of a habitual offender involved in 40 crimes, the system played a key role in locating the suspect.

Police have also found the system useful in tracking missing persons, as demonstrated when they used FRS to trace a techie from Bengaluru who had changed his appearance and was found in Noida.

Future of FRS in Bengaluru

As Bengaluru police continue to refine the technology, plans are in place to increase the number of FRS-linked cameras once the system’s accuracy improves. The system has shown potential, particularly in managing large crowds, where the FRS can monitor live feeds from political gatherings and festival celebrations to detect known criminals.

Highlights

FRS Functionality: AI-powered cameras capture live feeds and send them to the police command centre, where the facial recognition software compares the faces with the criminal database.
Goal: To stabilise the accuracy of the system to match the current standards of Automatic Number Plate Recognition (ANPR) cameras.
Impact: In 90 days, the system has identified 2.5 lakh people with criminal backgrounds and facilitated the arrest of 10 individuals.