Personnel
Overall Objectives
Research Program
Application Domains
Highlights of the Year
New Software and Platforms
New Results
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
Dissemination
Bibliography
XML PDF e-pub
PDF e-Pub


Section: New Results

Safe & Easy Environment for Alzheimer Disease and related disorders

Participants : Auriane Gros, François Brémond.

As part of the SafEE project (see 9.2.1.2), we have implemented two clinical protocols, one for patients in nursing home with Alzheimer disease and the other at home for patients with a frailty syndrome. The goal was to automatically recognize emotional disorders. In nursing home, we sought to detect events specific to the quality of sleep (effective hours of sleep, awakenings at night) to provide an objective measure of sleep disorders. For this purpose, several activity analysis methods have been used to increase the accuracy of automatic recognition. The events of interest were represented by "get in bed" and "get out bed". The performance of the three models used (A; B and B + C) allowed a performance of 0.71 (F1 score). At home, we focused on anxiety-like behavior (repetitive gestures, parasitic gestures), apathy (locomotion, gestures with goals) and agitation (retracing steps, locomotion). In order to be able to identify algorithms with the greatest precision possible, clinicians have realized annotations via the Viper software. In order to annotate these events we have developed a language based on a dedicated ontology for the recognition of activity behaviors that we have deemed of interest. In order to carry out an automatic recognition of the activities of interest, files (xml1 and xml3) were generated in the Scene Understanding Platform (SUP). For the recognition of the activity "to retrace one's steps" we have calculated the relation between two consecutive walks by comparing all the position points of these walks. The spatial properties information was extracted from the xml1 file and converted to Python Pandas data to analyze and visualize the data. The first results highlighted: -a gradual increase in social interactions during the follow-up; -an increase in gestures with goals; -a decrease in retracing one's steps. The studies were conducted on 6 patients during three months of follow-up in nursing home and three patients during six months of follow-up at home.