Image-Assisted Discrimination Method for Neurodevelopmental Disorders in Infants Based on Multi-feature Fusion and Ensemble Learning

Abstract

Premature infants have a significantly increased risk of developing severe neurodevelopmental disorders such as cerebral palsy and mental retardation due to some congenital defects at birth. During early infancy, distinct motion patterns occur which are highly predictive for later disability. The clinical observations of these forms of exercise can be record as parameters. In this paper, we used Kernel Correlation Filter (KCF) to track the trajectories of an infant’s limbs. Then, the obtained trajectories are analyzed in the wavelet domain and power spectrum domain, and integrated the features into the Ensemble Learning classification, the classification results are weighted and comprehensively judged to determine whether the infant’s neurodevelopment is normal and whether early intervention is needed.