Séminaire ARMEDIA présenté par Noé Zhang, Master student, EPH department, Telecom SudParis

Quand : Jeudi 24 juin 2021 à 14h00
Moyen : Webconf via le lien https://webconf.imt.fr/frontend/cat-clh-6al

Title : « Study of Facial Digital Markers for Early Detection of Parkinson’s Disease »

Abstract :
The primary goal of this study (led in collaboration with the Institut du Cerveau et de la Moelle épinière, ICM Paris) is to investigate the usefulness of digital markers extracted from facial videos in order to detect hypomimia symptoms (reduced facial dynamics) related to Parkinson disease. Although it is well known that hypomimia is an important Parkinsonian symptom, it has not yet been investigated largely in an automated way, on statistically significant databases, and respecting the privacy of the data. In order to achieve our goal, newly developed artificial intelligence methods (such as Deep Neural Nets), including transfer learning methods will be applied. The ICEBERG database, recorded in a medical context for our experiments, will be used.