SBIR Phase II: Monitoring and management of infant cranial development at the point-of-care

Pediametrix, Inc., Rockville MD

Investigators

Abstract

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase II project includes the support of newborn and infant health services at the point-of-care (i.e., in pediatrician offices and at home). This project will develop quantitative imagining and machine learning technology that will allow pediatricians to evaluate and monitor the infant’s head development during well-child visits and/or remotely. If a child is diagnosed with head deformations, early and effective therapy can be initiated. Thus, a major impact of this project is significant reduction of the number of children with cranial deformations who remain untreated and are exposed to potential health risks. Improved clinical management of these conditions will potentially lower the associated healthcare costs and improve patient outcome. Our technology will be packaged as a smartphone or tablet app, enabling access to lower-resourced communities and patients without access to specialized clinics, as well as enabling care in periods of social distancing and reduced clinical services. This project will ultimately increase awareness of the risks of head deformations and improve pediatric health outcomes. This Small Business Innovation Research (SBIR) Phase II project will develop quantitative imaging and machine learning algorithms that accurately measure the cranial shape and size using photos of the infant head taken with a smartphone. These parameters are commonly used to monitor child development, and in the diagnosis and treatment planning of common conditions with head deformations. In this project, we will fully automate the measurements using neural networks and maximize accuracy with regression models accounting for endogenous variables, such as sex and age. To generate accurate head shape measurements even during operation by novice users, our new algorithms based on machine learning will automatically select better images to detect and classify the type of cranial deformation. By addressing errors related to novice use of the tool, such as correcting for variations in the use of the camera, the technology will ensure compliance and accessibility to operators with variable technological skills. Furthermore, we will advance reconstruction of the three-dimensional cranial shape utilizing modern smart device technologies. This approach can handle object motion and can be implemented directly on a smartphone/tablet to compute full three-dimensional shape analysis. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.