New Delhi: Having maintained that over 2,800 people with “criminal antecedents” were part of the Cockroach Janta Party (CJP) protest in July and were “identified” using a facial recognition system (FRS), Delhi Police’s own officers say the accuracy of the system is between 60% and 80%, and if an image that is captured is unclear, it can return matches with similarity scores as low as 16% to 20%.Reports in The Print and The Indian Express have said that between 22 and 25 flagged by FRS during the protests were in jail at the time.Delhi Police officials Friday said they had not arrested any individual with criminal antecedents based on the FRS report and that an exercise was under way by local police stations to verify the whereabouts of the people flagged by the system.However, they did not respond to allegations that some people flagged by the system were in jail on July 20, when the Sansad Chalo march was called by CJP. The presence of each individual, police said, will be corroborated with prison details as well as mobile data locations, CCTV footage and witness testimony. Some may even be out on bail and still wrongly flagged, they said.Explaining how FRS works, cops said the software processes photographs and video footage against a pre-loaded database to identify potential suspects.Delhi Police has a database of more than 2.5 lakh people from across the country. “When investigators obtain a photograph or a clear frame from video footage, the image is fed into the system and compared with faces stored in the database. The software then analyses the facial features and generates potential matches, which can provide investigators with leads about the identity of a suspect,” officers said.The system analyses several facial parameters and assigns a similarity score to the possible matches. Officers set varying accuracy thresholds, such as 80% or 60%, but an FRS match serves strictly as an investigative lead rather than conclusive court evidence, police said.The officers said the system can generate up to 12 potential suspects or matches for a single image. In some cases, particularly when the source image is unclear or captured under difficult conditions, the system can return matches with similarity scores as low as 16% to 20%.Delhi Police uses software platforms developed by CDAC, Innefu and Fortify to aid field investigations. The officers said every match generated by the system is required to be independently verified by police personnel through physical verification, phone locations and other evidence.The quality of the source material is a major factor in the outcome. Clear, high-resolution photographs or video footage can sometimes produce matches with an accuracy of around 85%, though the system does not provide a 100% match. Images captured from a distance or with poor resolution and fewer pixels can produce less reliable results and a larger pool of potential matches.Delhi Police has been using vans fitted with FRS technology since 2018. Such vans were also placed at Jantar Mantar during the CJP-led protest.Each van is equipped with four cameras mounted on its roof, which capture images of people approaching or passing the vehicle. If the system detects a person whose face matches that of an individual in the criminal database, it immediately generates an alert. Local police personnel are then notified and tasked with verifying the person’s identity and antecedents.Officials said the technology is particularly useful in crowded areas and during major deployments, where manually identifying individuals from hundreds of people can be difficult. It has also played a major role in tracing missing people and children, police said.
