Automatic Quantification of the Degree of Flexion on Upper Extremity through Analysis of Monocular Video

Abstract

A method is introduced to obtain information on the movement of a human upper extremity based on automatic video analysis. This study used basic techniques of image segmentation and contour curvature calculation of the segmented extremity, as well as temporary dynamic alignment of sequences to track the two critical points that mark the elbow. Additionally, the median axes of the arm and forearm were calculated to then obtain curves of the extremity’s flexion angle over time. This methodology was used with seven individuals and the proposed automatic measurement was compared to the semi-assisted measurement in which a physiatrist conducted measurements on images by using a computer. The results suggest that this methodology could serve as a support tool in medicine for the evaluation of different diseases that affect the mobility of the extremities.