Early-stage parkinson's disease detection based on action unit derivatives - Archive ouverte HAL
Communication Dans Un Congrès Année : 2023

Early-stage parkinson's disease detection based on action unit derivatives

Anas Filali Razzouki
  • Fonction : Auteur
  • PersonId : 1288248
Laetitia Jeancolas
  • Fonction : Auteur
  • PersonId : 1288249
Graziella Mangone
  • Fonction : Auteur
  • PersonId : 1288250
Stéphane Lehéricy
Jean Christophe Corvol
Marie Vidailhet
Isabelle Arnulf

Résumé

Background: Hypomimia is a symptom of Parkinson’s disease (PD), characterizedby a decrease in facial movements and loss of face emotional expressions. This studyaims to detect hypomimia in participants with early-stage PD based on facial actionunits (AUs). Methods: A total of 299 video recordings were included, consisting of 208PD subjects and 91 healthy controls (HC), asked to perform fast syllable repetitions.To distinguish typical facial muscle movements from PD subjects associated withhypomimia, we compute the AUs derivatives. Global features were extracted basedon the AUs intensities and their derivatives, and XGBoost was used to classify PD vs.HC. Results: We obtain classification scores up to 73.00% in terms of balanced accuracy(BA) and an area under the curve (AUC) of 78.38% at video visit level. These resultsare promising for detecting hypomimia at an early stage of PD, and this work couldpotentially allow for continuous monitoring of hypomimia outside of hospitals throughtelemedicine.
Fichier principal
Vignette du fichier
Jetsan_2023_soumis_22_05.pdf (547.55 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04220902 , version 1 (28-09-2023)

Identifiants

  • HAL Id : hal-04220902 , version 1

Citer

Anas Filali Razzouki, Laetitia Jeancolas, Graziella Mangone, Stéphane Lehéricy, Jean Christophe Corvol, et al.. Early-stage parkinson's disease detection based on action unit derivatives. Colloque en TéléSANté et dispositifs biomédicaux ( JETSAN), Université Paris 8, CNRS,, Jun 2023, Paris Saint Denis, France. ⟨hal-04220902⟩
199 Consultations
327 Téléchargements

Partager

More