Markerless (ML) motion capture that allow to acquire motion analysis' parameters without the use of complex instrumentation such as stereophotogrammetric systems and the requirement of controlled environment such as gait laboratories.. When markerless motion capture is considered, the advantage is represented by the fact that single or multiple commercial video cameras can be used including mobile phone solutions. In this domain recently some approaches based on deep learning algorithms have been proposed (MediaPipe, OpenPose, Theia3D) . The main common applications of these technologies is the reconstruction of the three dimensional pose of the subject, the computation of the space temporal parameters describing the task or joint angles. An important aspect to be considered is the accuracy of these systems, and with this respect in the literature we can find some works based on RGB-D or other type of cameras. These works have shown that, at present, particularly reliable results can be obtained regarding spatiotemporal parameters, but there is still much to be done with regard to joint angles.
La tecnologia Markerless (ML) consente di acquisire parametri per l’analisi del movimento senza l'uso di strumentazioni complesse come i sistemi stereofotogrammetrici e la necessità di ambienti controllati come i laboratori di gait. Quando si considera il markerless, il vantaggio è rappresentato dal fatto che è possibile utilizzare videocamere commerciali singole o multiple, comprese le soluzioni di telefonia mobile. In questo ambito recentemente sono stati proposti alcuni approcci basati su algoritmi di deep learning (MediaPipe, OpenPose, Theia3D). Le principali applicazioni comuni di queste tecnologie sono la ricostruzione della posa tridimensionale del soggetto, il calcolo dei parametri spazio-temporali che descrivono il compito o gli angoli articolari. Un aspetto importante da considerare è la precisione di questi sistemi, e a questo proposito in letteratura possiamo trovare alcuni lavori basati su telecamere RGB-D o di altro tipo. Questi lavori hanno dimostrato che, attualmente, si possono ottenere risultati particolarmente affidabili per quanto riguarda i parametri spazio-temporali, ma c'è ancora molto da fare per quanto riguarda gli angoli articolari.
Ai driven markerless gait analysis
PENZO, FILIPPO
2023/2024
Abstract
Markerless (ML) motion capture that allow to acquire motion analysis' parameters without the use of complex instrumentation such as stereophotogrammetric systems and the requirement of controlled environment such as gait laboratories.. When markerless motion capture is considered, the advantage is represented by the fact that single or multiple commercial video cameras can be used including mobile phone solutions. In this domain recently some approaches based on deep learning algorithms have been proposed (MediaPipe, OpenPose, Theia3D) . The main common applications of these technologies is the reconstruction of the three dimensional pose of the subject, the computation of the space temporal parameters describing the task or joint angles. An important aspect to be considered is the accuracy of these systems, and with this respect in the literature we can find some works based on RGB-D or other type of cameras. These works have shown that, at present, particularly reliable results can be obtained regarding spatiotemporal parameters, but there is still much to be done with regard to joint angles.File | Dimensione | Formato | |
---|---|---|---|
Penzo_Filippo.pdf
embargo fino al 23/09/2027
Dimensione
2.38 MB
Formato
Adobe PDF
|
2.38 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/71135