Babies’ fingerprints may hold the key to a lifetime of better health. In a first-of-its-kind effort, Spartan researcher Anil Jain and his team are using biometric technology to collect the fingerprints of infants in places that lack consistent health and identification records with the goal of tracking immunizations and treatments from birth.
Taking fingerprints from very young children – even newborns – is part of a drive in developing countries to monitor the health of infants, who often lack other forms of identification. A biometric system, such as a national fingerprint database, could allow clinicians to match a child with their vaccine schedule or help workers keep records of welfare services, says Jain. “Vaccinations are first given at about 1 month, so that’s when we would like to use biometrics for recognition purposes,” he says.
“Taking babies’ fingerprints could help monitor vaccinations and identify infants swapped at birth“
Jain also thinks that taking infants’ fingerprints could help find missing children or resolve cases in which newborns are accidentally swapped at birth. Such cases are rare. But just this month Texas couple Richard Cushworth and Mercedes Casanellas were finally allowed to return home with their son more than a year after he was mixed up with another baby in a hospital in El Salvador. The legal process took so long partly because footprints taken at birth could not conclusively prove the babies’ identities and DNA tests were required.
Jain’s team worked with manufacturers of fingerprint scanners to build one that worked on infants. Tiny fingers have a denser pattern of ridges and valleys, so the device has to scan at a higher resolution. Even so, it can be tough to get a clear image from wriggling subjects with softer and more elastic skin. The team developed a machine-learning algorithm to enhance the scans.
The researchers tested the device at Saran Ashram Hospital in Agra, India. They visited the hospital four times over the course of a year and fingerprinted 319 babies. On follow-up visits, the team was able to identify infants first fingerprinted at 6 months or older with nearly 99 per cent accuracy.
This dropped significantly for younger children, however. For infants fingerprinted at a month or younger, the system was accurate less than half the time. Jain hopes to refine the method. The team presented the work at the International Conference on Information & Communication Technologies and Development in Ann Arbor, Michigan, this month.