Building a gait analysis framework for Parkinson#039;s disease patients: Motion capture and skeleton 3D representation

摘要

Stereoscopic vision has been very successful in recording and analyzing human motion. Using stereoscopic vision techniques and motion tracking of selected body joints, we have extracted and processed skeletal data to measure gait-related parameters and metrics of people with specific kinetic deficiencies. We have focused on Parkinson's Disease (PD) patients and have secured accurate representations of the kinesiological state of a patient at a given point in time (a kinesiological “imprint”). This stored imprint, for a given recording session, consists of a 3D skeletal visualization and associated derived gait metrics such as walking speed and stride length. In due time, a time sequence of accurate kinesiological imprints for any given patient will provide attending neurologists with evolutionary data of sufficient accuracy and granularity to better assess disease progress as well as quantify that patient's response to specific medications.