Understanding ASTR Facial Recognition

The Department of Telecommunications (DoT) takes pride in its innovative tool, the Artificial Intelligence and Facial Recognition powered Solution for Telecom SIM Subscriber Verification (ASTR) . This tool is designed to combat cybercrimes by identifying and obstructing potential fraudulent mobile connections.

• In 2012, a mandate was issued by the Department of Telecommunications (DoT) that required all telecom operators to share their subscriber database, including users’ photographs.
• The conception and design of the ASTR project took place between April 2021 and July 2021 under the guidance of the DoT’s unit.
• A pilot project was initiated in the Mewat region of Haryana to test the effectiveness and practicality of ASTR.
  • Prior to the ASTR pilot project, Mewat had roughly 16.69 lakh SIMs, out of which nearly 5 lakh SIMs across various telecom operators were identified as fraudulent.

Functioning of ASTR

• ASTR conducts a face comparison for each face against all faces in the database , grouping similar faces under one directory.
• ASTR concludes two faces to be identical if they match to the extent of at least 97.5 per cent . According to Jakhar’s paper:
  • ASTR can identify all SIMs associated with a suspected face in less than 10 seconds from a database of 1 crore images.
• Once faces are matched, ASTR's algorithm uses "fuzzy logic" to find similarities or approximate matches for the subscriber names.
• The DoT permits an individual to possess nine legitimate mobile phone connections using a single identity proof.
• The ASTR
  • Checks if there are more than nine connections against a single individual’s photograph;
  • It sifts through the database to find if the same person has acquired SIMs under different names.
• ASTR detected as many as 6,800 connections against a single image of the subscriber, that is, the same face, but different names. In another instance, 5,300 connections were obtained using a single image.

What's Next?

• Once the DoT identifies a set of numbers acquired through fraudulent means, it shares a list of these connections with telecom operators to block.
• An analysis of over 87 crore mobile connections was conducted using ASTR in the first phase, where more than 40 lakh cases of people using a single photograph to obtain mobile connections were detected.
  • After thorough verification, more than 36 lakh connections were disconnected by telecom operators.
• The same list is also shared with banks, payment wallets, and social media platforms to disengage these numbers from their respective platforms.