Investigating Fidgety Movement detection using consumer-based hardware and Artificial Neural Networks

Cerebral Palsy [CP] occurs before or at birth and frequently can be confounded with normal but slow development, therefore CP may go undetected for up to two years. If detected early, interventions can be implemented which reduce the lifetime impact for the subject. The early intervention begins the better the outcome. This thesis describes the process of developing an automatic classification system to look for signs of Cerebral Palsy. Specifically, the system examines videos of babies for the presence or absence of fidgety movements. The absence of fidgety Movement is a very strong indicator of Cerebral Palsy.