Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept Study

Abstract BackgroundMarkerless motion tracking methods have promise for use in a range of domains, including clinical settings where traditional marker-based systems for human pose estimation are not feasible. Artificial intelligence (AI)–based systems can offer a markerless, l...

Full description

Saved in:
Bibliographic Details
Main Authors: Vaidehi Wagh, Matthew W Scott, Sarah N Kraeutner
Format: Article
Language:English
Published: JMIR Publications 2024-12-01
Series:JMIR Formative Research
Online Access:https://formative.jmir.org/2024/1/e56682
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846100219535556608
author Vaidehi Wagh
Matthew W Scott
Sarah N Kraeutner
author_facet Vaidehi Wagh
Matthew W Scott
Sarah N Kraeutner
author_sort Vaidehi Wagh
collection DOAJ
description Abstract BackgroundMarkerless motion tracking methods have promise for use in a range of domains, including clinical settings where traditional marker-based systems for human pose estimation are not feasible. Artificial intelligence (AI)–based systems can offer a markerless, lightweight approach to motion capture. However, the accuracy of such systems, such as MediaPipe, for tracking fine upper limb movements involving the hand has not been explored. ObjectiveThe aim of this study is to evaluate the 2D accuracy of MediaPipe against a known standard. MethodsParticipants (N=10) performed a touchscreen-based shape-tracing task requiring them to trace the trajectory of a moving cursor using their index finger. Cursor trajectories created a reoccurring or random shape at 5 different speeds (500-2500 ms, in increments of 500 ms). Movement trajectories on each trial were simultaneously captured by the touchscreen and a separate video camera. Movement coordinates for each trial were extracted from the touchscreen and compared to those predicted by MediaPipe. Specifically, following resampling, normalization, and Procrustes transformations, root-mean-squared error (RMSE; primary outcome measure) was calculated between predicted coordinates and those generated by the touchscreen computer. ResultsAlthough there was some size distortion in the frame-by-frame estimates predicted by MediaPipe, shapes were similar between the 2 methods and transformations improved the general overlap and similarity of the shapes. The resultant mean RMSE between predicted coordinates and those generated by the touchscreen was 0.28 (SD 0.06) normalized px. Equivalence testing revealed that accuracy differed between MediaPipe and the touchscreen, but that the true difference was between 0 and 0.30 normalized px (t114Pt35.43P ConclusionsOverall, we quantified similarities between one AI-based approach to motion capture and a known standard for tracking fine upper limb movements, informing applications of such systems in domains such as clinical and research settings. Future work should address accuracy in 3 dimensions to further validate the use of AI-based systems, including MediaPipe, in such domains.
format Article
id doaj-art-0a5c77a3833f4be48fa1567dcbd924ff
institution Kabale University
issn 2561-326X
language English
publishDate 2024-12-01
publisher JMIR Publications
record_format Article
series JMIR Formative Research
spelling doaj-art-0a5c77a3833f4be48fa1567dcbd924ff2024-12-30T11:58:26ZengJMIR PublicationsJMIR Formative Research2561-326X2024-12-018e56682e5668210.2196/56682Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept StudyVaidehi Waghhttp://orcid.org/0009-0009-4813-1097Matthew W Scotthttp://orcid.org/0000-0003-1062-3490Sarah N Kraeutnerhttp://orcid.org/0000-0002-6552-6682 Abstract BackgroundMarkerless motion tracking methods have promise for use in a range of domains, including clinical settings where traditional marker-based systems for human pose estimation are not feasible. Artificial intelligence (AI)–based systems can offer a markerless, lightweight approach to motion capture. However, the accuracy of such systems, such as MediaPipe, for tracking fine upper limb movements involving the hand has not been explored. ObjectiveThe aim of this study is to evaluate the 2D accuracy of MediaPipe against a known standard. MethodsParticipants (N=10) performed a touchscreen-based shape-tracing task requiring them to trace the trajectory of a moving cursor using their index finger. Cursor trajectories created a reoccurring or random shape at 5 different speeds (500-2500 ms, in increments of 500 ms). Movement trajectories on each trial were simultaneously captured by the touchscreen and a separate video camera. Movement coordinates for each trial were extracted from the touchscreen and compared to those predicted by MediaPipe. Specifically, following resampling, normalization, and Procrustes transformations, root-mean-squared error (RMSE; primary outcome measure) was calculated between predicted coordinates and those generated by the touchscreen computer. ResultsAlthough there was some size distortion in the frame-by-frame estimates predicted by MediaPipe, shapes were similar between the 2 methods and transformations improved the general overlap and similarity of the shapes. The resultant mean RMSE between predicted coordinates and those generated by the touchscreen was 0.28 (SD 0.06) normalized px. Equivalence testing revealed that accuracy differed between MediaPipe and the touchscreen, but that the true difference was between 0 and 0.30 normalized px (t114Pt35.43P ConclusionsOverall, we quantified similarities between one AI-based approach to motion capture and a known standard for tracking fine upper limb movements, informing applications of such systems in domains such as clinical and research settings. Future work should address accuracy in 3 dimensions to further validate the use of AI-based systems, including MediaPipe, in such domains.https://formative.jmir.org/2024/1/e56682
spellingShingle Vaidehi Wagh
Matthew W Scott
Sarah N Kraeutner
Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept Study
JMIR Formative Research
title Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept Study
title_full Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept Study
title_fullStr Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept Study
title_full_unstemmed Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept Study
title_short Quantifying Similarities Between MediaPipe and a Known Standard to Address Issues in Tracking 2D Upper Limb Trajectories: Proof of Concept Study
title_sort quantifying similarities between mediapipe and a known standard to address issues in tracking 2d upper limb trajectories proof of concept study
url https://formative.jmir.org/2024/1/e56682
work_keys_str_mv AT vaidehiwagh quantifyingsimilaritiesbetweenmediapipeandaknownstandardtoaddressissuesintracking2dupperlimbtrajectoriesproofofconceptstudy
AT matthewwscott quantifyingsimilaritiesbetweenmediapipeandaknownstandardtoaddressissuesintracking2dupperlimbtrajectoriesproofofconceptstudy
AT sarahnkraeutner quantifyingsimilaritiesbetweenmediapipeandaknownstandardtoaddressissuesintracking2dupperlimbtrajectoriesproofofconceptstudy