Identifying Infants can be harder than it seems. Particularly in remote and limited resources settings, rapid and accurate identification of infants presents an unsolved complex sociotechnical problem. Imagine a long line of caregivers, each carrying several children, waiting outside in heat and humidity for required vaccinations. Caregivers may only know the infant's given names: how can the they be identified for record keeping?
Vaccination cards are notoriously unreliably and easily lost, mistakes abound. Recent technologycentered attempts th
In order to develop a new, infant-centered solution from the ground up, we assembled a diverse team of engineers, clinicians, ethnographers and designers and followed a Human Centered Design (HCD) approach of ethnography, rapid prototyping and testing. We examined all common modalities used in adult biometrics-- ear, iris, retina, face, foot, palm and finger recognition and compared technical feasibility, usability and acceptability for the infant use case. We prototyped many infant-centric devices and arrived at lead candidates using modified contact vs non contact palm and finger scanning. Frequent design-test cycles were critical as the complexity and changing nature of infant physiology, behavior and caregiver dynamics could not be predicted, only tested with subjects. This was compounded by moving targets of evolving infant-centric software, hardware and device design.
In summary, we report here an HCD based approach to infant biometrics. We developed and tested robust, socially acceptable technologies that adapt to the tiny, sensitive yet changing fingers of very young infants.