Bachelor Assignments

Listen up:
Unobtrusive voice sensing for elderly care in agitation

Status: Assigned
Student: Radu Seteanu

https://cdn.pixabay.com/photo/2015/07/23/11/37/sound-856771_960_720.png

Problem Statement: 

From the caregiver’s perspective, the need of unobtrusive voice sensing systems is prevalent for monitoring elderlies at a distance. Unobtrusive sensing broadly means a sensor-based system that doesn’t gain much of the user’s attention while detecting various required activities of the user. For quality elderly care, it is important to understand the emotional and cognitive state of elderlies. Interestingly, voice can be used to map subtle emotional and cognitive behavioral changes. Among others, agitation is a common neuropsychiatric disorder in elderlies and contributes to disability, institutionalization, and diminished quality of life for patients and their caregivers. Though agitation exhibits physical, cognitive and emotional behavior changes but our interest is voice activities like verbal aggression (screaming, shouting), and constant unwarranted requests for attention or help.

Tasks: 

With this project, we aim to detect the voice activity which can further give insight into current mood (may be level of agitation) of the elderlies. We expect to extract features of voice (for example pitch, prosody, loudness etc.) for reliable agitation monitoring and hence develop a basic framework which can distinguish levels of verbal aggression.


Work: 

20% Theory, 60% Simulations, 20%Writing.


Contact:  

Nikita Sharma (n.sharma@utwente.nl)