Classification of head movement patterns to aid patients undergoing home-based cervical spine rehabilitation

Abstract

Physical rehabilitation under close supervision of experts is often recommended to patients suffering from cervical pain and join related injuries. However, complete supervised rehabilitation may not be possible due to various socio-economical parameters. On the other hand, unsupervised rehabilitation may lead to post-injury complications. In this paper, we propose a pattern classification based method to understand head movement that is necessary to recognize unusual patterns during cervical spine rehabilitation. The proposed system takes the help of Kalman filter to fuse data acquired through camera and motion sensors. Patterns of the fused signals represent head displacement during left-right and front-back movements with respect to the central axes parallel to coronal and sagittal planes. Normal and abnormal patterns of movements are thereby represented using time series data that describes temporal change in the angle of head on above planes. Experimental validation has been done with data collected from several users. It has been observed that, our proposed methodology can precisely detect and represent abnormalities in head movements during cervical spine rehabilitation.