Bharat ABIS Aims to Tackle India's Duplicate Aadhaar Problem

Duplicate Aadhaar records have been a long-standing challenge for India’s identity system. Now, researchers associated with the Unique Identification Authority of India (UIDAI) have developed an open-source biometric verification system that could help detect duplicate identities more efficiently.

Called Bharat ABIS, the system has been tested against a gallery containing 220 million Aadhaar records. According to the researchers, the technology has demonstrated strong accuracy while also being designed to work with extremely large biometric databases.

Biometric identification is generally considered a highly secure method of verifying an individual. However, operating such a system at the scale of Aadhaar, which covers more than 1.5 billion people, presents significant technical challenges. The researchers behind Bharat ABIS say their system is designed to address some of these challenges while maintaining both speed and accuracy.

The research team comprises six UIDAI researchers, one researcher from IIIT Hyderabad and renowned biometrics researcher Anil Jain from Michigan State University. Their study, titled “Towards Billion-Scale Multi-modal Biometric Search,” examines how Bharat ABIS can be used to search large databases and identify potential duplicate identities.

The researchers specifically looked at how the system could add new records to a massive database without creating unnecessary duplicate entries or incorrectly flagging legitimate identities.

The issue is not new. In 2022, the Ministry of Electronics and Information Technology (MeitY) informed the Lok Sabha that nearly 6,00,000 duplicate Aadhaar cards had been identified and cancelled by May that year.

What is Bharat ABIS?

ABIS stands for Automated Biometric Identification System. Bharat ABIS combines three biometric identifiers - fingerprints, iris scans and facial information - into a single template measuring 13.5KB.

The resulting template can then be compared with records already stored in a database to determine whether the individual may already have an existing identity.

The system also includes a scheduler that distributes search tasks across multiple servers, an important feature when dealing with databases containing hundreds of millions or potentially billions of records.

Bharat ABIS can work with individual biometric identifiers, such as fingerprints or iris scans. However, the researchers found that combining all three biometric layers significantly improves accuracy.

How accurate is Bharat ABIS?

According to the study, Bharat ABIS achieved a False Positive Identification Rate (FPIR) of 0.1 per cent. FPIR represents instances where the system could incorrectly identify an individual as a duplicate. UIDAI’s target rate for vendors is 0.5 per cent.

The system recorded a False Negative Identification Rate (FNIR) of 0.05 per cent, referring to cases where a duplicate could be missed. The target set for this measure is 0.1 per cent. The researchers note that both figures apply to a 20 million-record database.

The team also compared Bharat ABIS with three existing tools used within the Aadhaar ecosystem. On the 20 million-record database, the researchers found that the open-source system delivered “comparable or better” accuracy than all three tools.

Another potential benefit comes from its DeepPrint fingerprint model. When combined with vendor tools, the technology can reduce authentication failures by half, according to the study. The researchers said the biggest improvements were seen among older people and manual labourers, whose fingerprints can become worn over time.

In terms of speed, Bharat ABIS can reportedly perform 100 searches per second on a 40 million-record gallery using a single server equipped with eight Nvidia H100 GPUs and 2TB of RAM.

The researchers also believe the open-source system could be useful in around 160 countries with populations below 20 million. Such countries may find commercial biometric systems expensive, while an open-source platform could provide an alternative.

However, Bharat ABIS has not yet been conclusively tested at the full scale of the Aadhaar database. The researchers acknowledge that further large-scale testing is necessary.

For now, Bharat ABIS appears to offer a promising approach to large-scale biometric identification, but more testing will be needed before its suitability for a system as large as Aadhaar can be established.

ABOUT THE AUTHOR

Kahekashan

[email protected]

Kahekashan is a passionate technophile with a keen eye for cutting-edge gadgets, emerging technologies, and everything in the digital realm. Raised in a Defence family with strong values and a background in literature, she has consistently pursued excellence in every endeavour. Her last full-time assignment involved content writing with the Indian School of Business.

Read More