Assessment of kinematic cmj data using a deep learning algorithm-based markerless motion capture system

The purpose of this study was to compare the performance of a 2D video-based markerless motion capture system to a conventional marker-based approach during a counter movement jump (CMJ). Twenty-three healthy participants performed CMJ while data were collected simultaneously via a marker-based (Oqus) and a 2D video-based motion capture system (Miqus, both: Qualisys AB, Gothenburg, Sweden). The 2D video data was further processed using Theia3D (Theia Markerless Inc.), both sets of data were analysed concurrently in Visual3D (C-motion, Inc). Excellent agreement between systems with ICCs >0.988 exists for Jump height (mean average error of 0.35 cm) and ankle and knee sagittal plane angles (RMS differences < 5°). The hip joint showed higher.
© Copyright 2020 ISBS Proceedings Archive (Michigan). Northern Michigan University. Published by International Society of Biomechanics in Sports. All rights reserved.

Bibliographic Details
Subjects:
Notations:training science biological and medical sciences technical and natural sciences
Tagging:deep learning Algorithmus reaktiver Sprung markerless
Published in:ISBS Proceedings Archive (Michigan)
Language:English
Published: Liverpool International Society of Biomechanics in Sports 2020
Online Access:https://commons.nmu.edu/isbs/vol38/iss1/218
Volume:38
Issue:1
Pages:Article 218
Document types:congress proceedings
Level:advanced