Delhi Police Seeks Facial Recognition CCTV Cameras for Public Spaces

Delhi Police has initiated a tender process to acquire 338 CCTV cameras featuring facial recognition technology. The purchase will focus on critical areas, including Ramlila Maidan, and is intended to bolster security and surveillance efficiency, particularly during large gatherings and along routes used by VIPs.

The request for proposal specifies that the cameras must include advanced analytics for crowd detection, identification of abandoned objects, and vehicle license plate recognition. Both fixed and pan-tilt-zoom (PTZ) cameras are required to support facial recognition, and automatic number plate recognition (ANPR) cameras are also expected to be capable of face detection.

The estimated value of the tender is Rs 15.66 crore (approximately $1.9 million USD). The bidding process closed on August 20. Concerns have been raised regarding privacy, as India lacks specific legislation governing the police use of facial recognition. The tender document mentions privacy only in the context of "privacy masking," a feature that allows for blurring or blacking out specific areas or faces in video footage.

Notably, the tender sets no accuracy standards for the facial recognition system or requirements for facial databases. Critics argue that if privacy safeguards are not prioritized at the product specification stage, the resulting systems are unlikely to uphold privacy-preserving principles. Previous reports indicated that Delhi Police's facial recognition system identified over 2,800 individuals at a protest site, though subsequent findings revealed inaccuracies in these identifications. The cameras are also expected to integrate with the Delhi Police's Safe City project, which includes predictive policing tools.