Two-dimensional video-based analysis of human gait using pose estimation

Abstract

Walking is the primary mode of human locomotion. Accordingly, people have been interested in studying human gait since at least the fourth century BC. Human gait analysis is now common in many fields of clinical and basic research, but gold standard approaches – e.g., three-dimensional motion capture, instrumented mats or footwear, and wearables – are often expensive, immobile, data-limited, and/or require specialized equipment or expertise for operation. Recent advances in video-based pose estimation have suggested exciting potential for analyzing human gait using only two-dimensional video inputs collected from readily accessible devices (e.g., smartphones, tablets). However, we currently lack: 1) data about the accuracy of video-based pose estimation approaches for human gait analysis relative to gold standard measurement techniques and 2) an available workflow for performing human gait analysis via video-based pose estimation. In this study, we compared a large set of spatiotemporal and sagittal kinematic gait parameters as measured by OpenPose (a freely available algorithm for video-based human pose estimation) and three-dimensional motion capture from trials where healthy adults walked overground. We found that OpenPose performed well in estimating many gait parameters (e.g., step time, step length, sagittal hip and knee angles) while some (e.g., double support time, sagittal ankle angles) were less accurate. We observed that mean values for individual participants – as are often of primary interest in clinical settings – were more accurate than individual step-by-step measurements. We also provide a workflow for users to perform their own gait analyses and offer suggestions and considerations for future approaches.

Video-based quantification of human movement frequency using pose estimation

“…The pros and cons of using currently available pose estimation algorithms for measurement of human movement have been discussed at length (29), and we will discuss our own impressions and suggestions below. Despite current limitations, this study and our prior work (21) have shown promising results when comparing pose estimation outputs to groundtruth measurements of clinically relevant movement parameters. We anticipate that the accuracy of human pose estimation algorithms will continue to improve, and we offer that this emerging technology has significant potential for versatile, quantitative, remote assessment of human motor function.…”

“…While emerging evidence suggests significant potential for tracking repetitive movements using video-based pose estimation approaches, here we addressed two important unmet needs: 1) testing across a range of clinically relevant tasks spanning the upper and lower extremities, and 2) assessment of accuracy via validation against ground-truth measurement. Based on our prior work (21), we expected good agreement between the OpenPose estimates of movement frequency and ground-truth measurement. We also share our perspectives regarding limitations, potential problems, and suggestions for successful analysis.…”

“…In this study, we used OpenPose (a freely available human pose estimation algorithm (11,15,16) to measure the frequencies of repetitive upper and lower extremity movements using smartphone videos. Previous studies have used OpenPose to study human gait (17)(18)(19)(20)(21) and quantify tremor severity in persons with PD (22). Other pose estimation algorithms have been used to measure the degree of levodopa-induced dyskinesias (23) and finger tapping frequency in PD that correlated well with clinical ratings (24,25).…”

Technical Note: Quantifying music-dance synchrony with the application of a deep learning-based 2D pose estimator

“…OpenPose has already been used in gait analysis research. 34,44,45 We believe that our work here adds to the research on using accessible and affordable deep-learning-based keypoint estimators in the analysis of complex movements such as dance.…”

Automatic Ankle Angle Detection by Integrated RGB and Depth Camera System

“…Erika D’Antonio (2020) [ 34 ] used OpenPose to investigate the accuracy of such a system estimating kinematic parameters of human gait using two webcams synchronized with inconclusive results, commenting on the need for further analysis. Stenum (2020) [ 35 ] used a system based on RGB images recording a person from the side while walking to obtain gait parameters employing OpenPose, with high accuracy of 0.89 for ankle angle.…”