NEW DELHI: The Central government has announced that it has given its approval for the implementation of the Automated Facial Recognition System (AFRS). This system, once implemented across the country, will be able to, among other things, extract an image from a video and match it with the image of an individual whose record is already in an existing database. Once the AFRS goes online, sometime early next year, it will be the world’s largest facial recognition system.

While speaking in Parliament, Union Minister of State for Home Affairs, G. Kishan Reddy on Wednesday said that permission to implement the system, which will change the way policing is done in India, has been approved by the government.

The bid for purchase of the system, which will be handled by the National Crime Records Bureau (NRCB), under the MHA, was first put out in the public domain on 3 July last year and it was supposed to close on 16 August. However, so far, the bid-opening date has been extended eight times and now the new date for the opening of the bid is 30 March.

The AFRS will be installed at the NCRB headquarters in Mahipalpur.

By announcing that it will be implementing the system, the government has bypassed all privacy concerns that were being raised by groups and individuals who fear that once the system is implemented in totality, the movement, location of any individual will be added into the records of the government without his or her consent being sought for the same.

The 172-page Request for Proposal (RPF) document or the tender specification documents released by the NCRB, seeking bid for the AFRS, gives an insight into how the government plans to use the system with the intent to prevent crime and identify criminals.

The NRCB, in the first few pages of the RPF, has stated that the “Automated Facial Recognition System” can play a very vital role in improving outcomes in the area of criminal identification and verification by facilitating easy recording, analysis, retrieval and sharing of information between different organizations.

“The Automated Facial Recognition System (AFRS) would help in automatic identification and verification of persons from digital images, photos, digital sketches, video frames and video sources by comparison of selected facial features of the image from an already existing image database. A Facial Recognition System is a great investigation enhancer for identification of criminals, missing children/persons, unidentified dead bodies and unknown traced children/persons.”

The use of AFRS, once functional, will be available to both state police as well as the Central forces and agencies that come under the Central government.

The NCRB, stating the benefits of the system, has said that it will work as a repository of photographs of criminals in the country and will give enhanced ability to detect crime patterns and understand the modus operandi of criminals across the states.

It will aid in crime prevention as police personnel can check the suspect with the hotlist of criminals by clicking the picture of the suspect on their mobile phone which will be then cross checked with the AFRS data-base at Mahipalpur. The result of this will be available within three minutes.

It is envisaged that apart from the Central data repository in Mahipalpur, there will be state level repository where crime and criminal related facial image data will be stored.

The AFRS will use/capture data from multiple sources including images from CCTV feed and generate alerts if a blacklisted match is found. It will also add photographs to its database obtained from newspapers, police raids, pictures sent by people and police sketches. It will also be able to add details like sex, age, scars, tattoos, etc., for future searches.

“It will be able to search photographs from the database, matching with the suspect features. It will be able to ‘lift’ photograph images available with Passport, CCTNS, ICJS and Prisons, Ministry of women and child development (KhoyaPaya) State or National Automated Fingerprint Identification System or any other image database available with police/other entity,” the document reads.

The AFRS will also match suspected criminal face from pre-recorded video feeds obtained from CCTVs deployed in various critical identified locations, or with the video feeds received from private or other public organization’s video feeds.

“It will be of a great help in a riot like situation to identify rioters and those who are in the mob. Once CCTV records their faces, the images from the video will be uploaded to the AFRS and within minutes it will be able to identify the suspect, as it will match the image from the multiple database it already has access to”, an official source explained. According to him, at the state level, the authorised requests for search, matching and verification of facial image would come from police stations. The facial recognition system shall be enabled at cameras identified by the Authority. These cameras identified shall be installed at critical locations finalised by the Authority.

The system that the MHA is looking for will be able to recognise the individual “regardless of vantage point and facial changes (glasses, beard, and expression)”. Hence, disguise may not work in concealing the identity of the suspect.

The MHA also intends to provide handheld mobiles with an application that will be connected to the AFRS to capture a face on the field and get the matching result from the backend server.

The AFRS is expected to handle at least 2,500 concurrent users.

Explaining the process as to where the data for the AFRS will be collected from, the official said that data will be sourced from police stations, crime records and from surveillance cameras located at different locations in the cities.

While responding to a legal notice sent by the Delhi-based Internet Freedom Foundation, a non-governmental organization that conducts advocacy on digital rights and liberties, in which the Foundation has raised concerns regarding violation of privacy and other allied issues, the NCRB said that it will be using data from Crime and Criminal Tracking Networks and Systems (CCTNS) which is secure and not available in the public domain. However, the RPF clearly states that the AFRS will be using data from multiple sources and not just CCTNS.

The NCRB in its reply further stated that AFRS will automate the matching process of suspects with the police records which is presently done manually. “AFRS automates this matching process and provides a bigger set for comparison, as it will run on state/nation level CCTNS/ICJS database. AFRS results will be further corroborated and analyzed by collecting other evidence by IO before drawing any conclusion. The AFRS will not source facial images from CCTV cameras in public places, unless the video footage is part of scene of crime”. The NCRB added that it does not plan to integrate AFRS with the Aadhaar database and will not use the same for “discrimination profiling”.