Please use this identifier to cite or link to this item:
Title: Easing automatic neurorehabilitation via classification and smoothness analysis
Authors: Bensalah, Asma
Fornés, Alicia
Carmona-Duarte, Cristina 
Lladós, Josep
UNESCO Clasification: 1203 Ciencia de los ordenadores
120304 Inteligencia artificial
Keywords: Deep Learning
Movement Classification
Movement Smoothness
Neurorehabilitation, et al
Issue Date: 2022
Publisher: Springer 
Journal: Lecture Notes in Computer Science 
Conference: IGS2021: The 20th Conference of the International Graphonomics Society
Abstract: Assessing the quality of movements for post-stroke patients during the rehabilitation phase is vital given that there is no standard stroke rehabilitation plan for all the patients. In fact, it depends basically on the patient’s functional independence and its progress along the rehabilitation sessions. To tackle this challenge and make neurorehabilitation more agile, we propose an automatic assessment pipeline that starts by recognising patients’ movements by means of a shallow deep learning architecture, then measuring the movement quality using jerk measure and related measures. A particularity of this work is that the dataset used is clinically relevant, since it represents movements inspired from Fugl-Meyer a well common upper-limb clinical stroke assessment scale for stroke patients. We show that it is possible to detect the contrast between healthy and patients movements in terms of smoothness, besides achieving conclusions about the patients’ progress during the rehabilitation sessions that correspond to the clinicians’ findings about each case.
ISBN: 978-3-031-19744-4
ISSN: 0302-9743
DOI: 10.1007/978-3-031-19745-1_25
Source: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) [ISSN 0302-9743], v. 13424 LNCS, p. 336-348, (Enero 2022)
Appears in Collections:Actas de congresos
Show full item record

Page view(s)

checked on Feb 26, 2023

Google ScholarTM




Export metadata

Items in accedaCRIS are protected by copyright, with all rights reserved, unless otherwise indicated.