Computer Vision-Based Infant Movement Assessment

Abstract

The aim of this thesis is to develop an objective predictive model for cerebral palsy based on video recording of spontaneous movements of young infants. In the recent years, computer-based motion assessment methods have aimed to quantitatively analyze infants’ movements in order to predict CP. However, they either use extra instruments that are intrusive to the diagnosis task, or their results are clinically hard to interpret. Thus, in order to tackle the problems with previous methods and perform an analytic tool to study the disease, the study proposes a new motion assessment tool to separate and track different body parts individually and without need for any extra instrument. In addition, the research achieves a model for prediction of cerebral palsy based on motion data of young infants. The prediction is formulated as a classification problem to assign each of the infants to one of the healthy or with cerebral palsy group. The thesis studies the previously proposed features for predicting cerebral palsy and in addition proposes a new feature set. Unlike formerly proposed features that are mostly defined in the time domain, this study proposes a set of features derived from frequency analysis of infants’ motions. Since cerebral palsy affects the variability of the motions, suggested features are suitable and consistent with the nature of the condition. Finally, in the current application, a well-known problem, few subjects and many features, was initially encountered. In such a case, most classifiers get trapped in a sub-optimal model and, consequently, fail to provide sufficient prediction accuracy. To solve this problem, a feature selection method that determines features with significant predictive ability is proposed. The feature selection method decreases the risk of false discovery and, therefore, the prediction model is more likely to be valid and generalizable for future infants.