17 Jul 2010, 0326 hrs IST,Sruthijith KK,ET Bureau
NEW DELHI: The logo of Aadhaar, the government’s unique identification project headed by technocrat Nandan Nilekani, shows a part of a thumb impression — a sharp, curving pattern found on the edge of the index finger. In the complex task of uniquely identifying 1.14-billion people that populate this country, the unique patterns on the fingers are Nilekani’s biggest allies.
But the Unique Identification Authority of India (UIDAI) the former Infosys boss heads is facing challenges from a strange quarter, challenges that go into the core of its very mission.
Scores of people the Aadhaar project will help the most do not have the sharp, curving lines on their fingers as depicted in its logo. Millions of Indians working in agriculture, construction workers and other manual labourers have worn-out fingers due to a lifetime of hard labour, resulting in what is euphemistically referred to in technical literature as ‘low-quality’ fingerprints. This is precisely the demographic that UID aims to help — those that are outside government records and welfare schemes.
While the UIDAI uses two other metrics — an iris scan and a photograph — in issuing the unique identity number, fingerprinting will be the metric used in authentication. This means a passport applicant with worn-out fingers may present his newly-issued UID number as a conclusive proof of identity, but could find the application rejected. The authentication process using a fingerprint scanner could classify the applicant’s worn-out fingers as a so-called ‘false negative’.
A December 2009 report by the UIDAI Committee on Biometrics, says there is no estimation of the extent of this problem. “The fingerprint quality, the most important variable for determining accuracy, has not been studied in depth in the Indian context,” the report says.
Subsequently, a pilot study was done, and 250,000 fingerprints were collected and analysed. The committee’s conclusion: “There is good evidence to suggest that fingerprint data from rural India may be as good as elsewhere when proper operational procedures are followed and good quality devices are used ... (but) the quality drops precipitously if attention is not given to operational processes.”
In the pilot study, 2-5% of subjects were found to not have any biometric data. “Missing biometrics is a license to commit fraud,” the study notes.
In issuing a unique identification number, the system uses three metrics — a photograph, an iris scan and a fingerprint. The iris scan and the fingerprint are used for a process known as deduplication. When a person comes to a UID counter to get a number, the software can instantly run the fingerprint and iris scan against a database to see if the applicant has already been issued a number.
“We use the iris scan as a deduplication metric. Fingerprint is the authentication metric, apart from the photograph,” said RS Sharma, director general and mission director of UIDAI.
But even with the iris scan, there can be issues. An iris scan cannot be done on people with corneal blindness or corneal scars. There are an estimated 6-8 million people in India with corneal blindness, according to a 2005 paper by three researchers at the Rajendra Prasad Centre for Ophthalmic Sciences at the All India Institute of Medical Sciences, New Delhi. The number of people with corneal scars will be much more. Corneal scars result from injury or infection of the cornea.
Another issue is cataract, which results from nutritional deficiencies and prolonged exposure to sunlight and UV rays. According to Dr G Chandra Shekhar, director of the LV Prasad Eye Institute in Hyderabad, persons undergoing cataract surgery will almost certainly have their iris affected. An iris scan cannot be done on such persons. Experts estimate that up to one in 200 people get a cataract surgery every year. But the Unique Identification Authority of India (UIDAI) the former Infosys boss heads is facing challenges from a strange quarter, challenges that go into the core of its very mission.
Scores of people the Aadhaar project will help the most do not have the sharp, curving lines on their fingers as depicted in its logo. Millions of Indians working in agriculture, construction workers and other manual labourers have worn-out fingers due to a lifetime of hard labour, resulting in what is euphemistically referred to in technical literature as ‘low-quality’ fingerprints. This is precisely the demographic that UID aims to help — those that are outside government records and welfare schemes.
While the UIDAI uses two other metrics — an iris scan and a photograph — in issuing the unique identity number, fingerprinting will be the metric used in authentication. This means a passport applicant with worn-out fingers may present his newly-issued UID number as a conclusive proof of identity, but could find the application rejected. The authentication process using a fingerprint scanner could classify the applicant’s worn-out fingers as a so-called ‘false negative’.
A December 2009 report by the UIDAI Committee on Biometrics, says there is no estimation of the extent of this problem. “The fingerprint quality, the most important variable for determining accuracy, has not been studied in depth in the Indian context,” the report says.
Subsequently, a pilot study was done, and 250,000 fingerprints were collected and analysed. The committee’s conclusion: “There is good evidence to suggest that fingerprint data from rural India may be as good as elsewhere when proper operational procedures are followed and good quality devices are used ... (but) the quality drops precipitously if attention is not given to operational processes.”
In the pilot study, 2-5% of subjects were found to not have any biometric data. “Missing biometrics is a license to commit fraud,” the study notes.
In issuing a unique identification number, the system uses three metrics — a photograph, an iris scan and a fingerprint. The iris scan and the fingerprint are used for a process known as deduplication. When a person comes to a UID counter to get a number, the software can instantly run the fingerprint and iris scan against a database to see if the applicant has already been issued a number.
“We use the iris scan as a deduplication metric. Fingerprint is the authentication metric, apart from the photograph,” said RS Sharma, director general and mission director of UIDAI.
But even with the iris scan, there can be issues. An iris scan cannot be done on people with corneal blindness or corneal scars. There are an estimated 6-8 million people in India with corneal blindness, according to a 2005 paper by three researchers at the Rajendra Prasad Centre for Ophthalmic Sciences at the All India Institute of Medical Sciences, New Delhi. The number of people with corneal scars will be much more. Corneal scars result from injury or infection of the cornea.
This means that there could be a significant overlap between people with ‘low-quality’ fingerprints and those with corneal/cataract problems, especially because they are likely to be from a similar demographic. A farm labourer doing manual work is both likely to develop worn-out fingers and a cataract problem due to exposure to sunlight. Such persons carry a greater risk of injuring their eyes because of accidents.
The set of people in the intersection of both these conditions pose a unique problem to UIDAI. “We expect such a set to be a small number, but yes, we will have to rely only on photographs in such cases,” says Sharma at UIDAI. He said the agency had no estimate of how many people could potentially have such issues.
While Nilekani was not available for comment, Sharma is unfazed. “We are dealing with a large country and complex issues. We have to work within these limitations,” he said, adding that UIDAI’s software has provisions for dealing with people without sufficient biometrics. The system treats these as exceptions.
Other experts said these problems are not insurmountable. Samir Brahmachari, director general of the Council for Scientific and Industrial Research, said he has discussed this issue with UIDAI officials. “It’s a problem and I think they could use DNA fingerprinting in such cases. That’s unique and 100% accurate,” he said.
But Sharma is less sanguine about this particular solution, and argues that using a different metric for a section of the database would create more loopholes in the system. “We have to measure everyone by the same metric. Else, deduplication won’t be possible,” he said.
