You are reading it here first: Soon, the Bangalore City railway station will be inundated with a myriad of CCTV cameras, all capable of carrying out facial recognition surveillance — at the rate of at least one hundred faces per second, according to a tender document released by the South Western Railways earlier this month. The surveillance system would be able to generate audio-visual alerts upon spotting a person in a “watchlist”, and display their complete profile and case history from a criminal database module, per the document. It will also be able to classify faces on the basis of their sex and age.

The advertised cost of setting up this surveillance infrastructure is humongous — ₹4,50,49,056. The last date of submitting bids is November 11.

The system will be integrated with the Crime and Criminal Tracking Network and Systems (CCTNS), a centrally maintained database of criminals in India. To monitor alerts generated by the system, a command centre at the railway station will also be set up. Aside from just facial recognition, it will operate with access to an “unlimited” number of records in databases, and create “unlimited” number of lists such as a “black list, a white list, and a watch list”.

This development is in line with an increasing trend among Indian government bodies: from organisations that track crime statistics such as the NCRB, to  an education board like the CBSE, every institution is deploying facial recognition systems — all while India lacks a data protection law. The Indian Railways itself had earlier said it was installing facial recognition surveillance cameras at more than 900 railway stations, across the country.

For Bangalore’s railway station, two different kinds of surveillance cameras will share duties to carry out facial recognition — a 2 megapixel IP camera, and an 8 megapixel infrared camera. The IP cameras should have “excellent low light performance”, suggesting that they will be in use round the clock, the tender said. This camera should also be able to integrate with a criminal analysis software called Eagle Eye. It is not clear which company has built this software, and when the Railways procured it; we have reached out to them for details.

Match faces in bulk, integrate with CCTNS: How the surveillance tool looks

The surveillance tool has a broad mandate. It has purposes beyond facial matching including perimeter surveillance, people counting, and object detection, among other things. Other requirements:

  • The software must allow adding facial data in bulk, and should also be able to integrate images from the Crime and Criminal Tracking Network and Systems. It should also have an extensive search feature based on a person’s name, date range, and criminal ID, among other things.
  • It should be able to carry out crime scene mapping, and provide such mapping within a given jurisdiction.
  • The system should work on partial occlusion of face, glasses, scarf, change of facial expression etc.
  • The software must be capable of historical data search when a new photo is introduced to the system; it should be able to search from historical data if the match is found and the number of times it occurs in the database.
  • Fingerprint matching as well: The system should also be capable of creating a new fingerprint database, and integrate with existing databases.

High accuracy demand: The Railways’s demand for the facial recognition’s accuracy seem ambitious — only a failure rate of one in one million is acceptable, which means an accuracy rate of 99.9%. The algorithm should be certified by US’ National Institute of Standards and Technology, a fairly common requirement in most facial recognition tenders floated by the Indian government in recent years.

A mobile app to make facial recognition portable: Not just that, a mobile app, with facial recognition capabilities, will also have to be developed which will be used by Railway police officials. It should also receive any notifications, in real-time, should the larger surveillance system generate any alert. Twenty five officials should be able to use the app at the same time, the tender said.

  • The app will also be able to carry out live fingerprint matching of a person, by “securely communicating with the Biometric Services Gateway in the central server”, per the tender document. “The central server having the Fingerprint Database behind a secure firewall should analyze and return the fingerprint results to the officer’s smartphone on the spot”.

Everyone’s deploying a facial recognition tool

From several departments of Indian government itself to several state governments and their respective police departments, the use of facial recognition technology is proliferating across the country. Recently, CBSE rolled out a facial recognition system — sans a privacy policy — for students to download their academic documents. The National Crime Records Bureau is currently working towards building a national level facial recognition system, and only very recently revealed that it wants to test the system on mask-wearing faces, and for it to generate “comprehensive biometric reports”.

Several airports have added facial recognition systems as an additional way for passengers to board flights. Local police departments are purchasing, or piloting facial recognition algorithms — for instance, the Vadodara city police piloted Clearview AI’s controversial facial recognition system. States like Telangana have piloted the tech in civic elections, and are mulling introducing the tool for obtaining ration at fair price shops.

It is worth noting that Indian agencies are installing facial recognition surveillance tools when India doesn’t have a data protection law. Moreover, the Personal Data Protection Bill, 2019, which is currently being deliberated upon by a Joint Parliamentary Committee, has carved out exemptions for government agencies to adhere to provisions of the Bill. This suggests that government institutions like the NCRB, or the Railways could potentially be able to collect, store and process biometric data of Indians without necessarily adhering to provisions of the Bill.

More from our facial recognition coverage: