Billions of dollars a year are spent vaccinating children in developing countries, but about half as many immunizations are administered as could be because of unreliable vaccination records. Biometric researchers from Michigan State University have developed a fingerprint-scanning system for children under five years old that could replace ineffective paper vaccination records.
Until now, biometrics experts believed fingerprints of babies and toddlers were too unreliable because image sensors are designed for the ridges and valleys of adult fingertips. The Michigan State University researchers developed software that makes it feasible to accurately match fingerprints of children under five with off-the-shelf equipment. They intend to present a paper detailing their work at a biometrics conference later this month. Paper-based vaccination records are easily lost and don’t reliably provide health workers with up-to-date information on patient history. Fingerprints are a better biometric trait than the iris of the eye or palm and footprints because they are easier to record from young children and the sensors are small and work quickly.
The Bill and Melinda Gates Foundation funded the research specifically for recording vaccinations, but the patient-identifying system has broader applications, says Anil Jain, a distinguished professor at Michigan State University’s Computer Science and Engineering Department and coauthor of the paper. The technology could be used “in any health-care scenario where you have the potential for fraud, such as insurance fraud,” says Jain.
Recording fingerprints from infancy into early childhood and adulthood would also allow countries to link civil registries, such as birth and death certificates to health care records, school enrollments, and voter registration.
The Michigan State University researchers needed to process images taken from fingerprint sensors using software to compensate for the small size of the children’s fingerprints, as well as their sometimes wet and oily skin. They also improved accuracy by creating matches based on both thumbs and index fingers.
In a trial in Benin, West Africa, the software successfully matched about 70 percent, compared to 98 percent in another test in Lansing, Michigan. The lower accuracy is because the Benin clinic was outdoors, with dust and high humidity, but Jain believes match accuracy can be improved to 95 percent in these types of conditions. The Michigan State University researchers plan to do further tests, potentially in India, where already there is a national biometric ID program.
Privacy groups have warned that biometric data can be misused or faked. However, people in developing countries typically do not voice concerns over privacy and security with biometric data, says Mark Thomas, the executive director of VaxTrac, a nonprofit that worked with the Benin trials. “Compared to the status quo of not getting vaccinations for life-threatening diseases, mothers are not asking these questions because the benefits are so intuitive,” he says.
How AI is reinventing what computers are
Three key ways artificial intelligence is changing what it means to compute.
These weird virtual creatures evolve their bodies to solve problems
They show how intelligence and body plans are closely linked—and could unlock AI for robots.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
Surgeons have successfully tested a pig’s kidney in a human patient
The test, in a brain-dead patient, was very short but represents a milestone in the long quest to use animal organs in human transplants.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.