Envisioning Recovery after Brain Injury: Integrating Computer Vision into Clinical Assessment of the Upper Limb

Research Objectives: To determine feasibility of kinematic analysis of the upper extremity using camera footage of a standardized clinical assessment, which is commonly used to assess motor function in patients with brain injury.

Design: We present a case study involving repeated measures performed during a single session.

Setting: Study setting is a clinical research laboratory located within an inpatient rehabilitation facility.

Participants: We recruited a single neurotypical participant to perform a target of five repetitions of tasks.

Interventions: Given the preliminary nature of computer vision in this application, an intervention was not applicable to this case study.

Main Outcome Measures: Main outcome measures are upper extremity kinematics as determined via a conventional optoelectronic motion capture system and as determined by a computer vision algorithm applied to simultaneous video footage captured with two normal cameras.

Results: Based on activities performed during the Action Research Arm Test, average joint position and joint angles derived from Computer Vision are comparable to values obtained by a conventional optoelectronic motion capture system.

Conclusions: This study provides preliminary evidence supporting feasibility of computer vision to obtain upper extremity kinematics during the Action Research Arm Test. Integration of a two-camera video recording into standardized clinical assessment may facilitate creation of an expansive database of upper extremity kinematics in neurorehabilitation. This would support research on neural and behavioral restitution following neurologic diseases such as stroke and traumatic brain injury.

Author(s) Disclosures: None