University of Twente Student Theses


Emotional activity detection using behavioural sounds

Knol, R. (2020) Emotional activity detection using behavioural sounds.

This is the latest version of this item.

[img] PDF
Abstract:Emotion detection is a desirable feature for monitoring systems in healthcare. Many data sources have been examined, like facial expressions, physiological signals, and speech. A yet unused data source are sounds produced by human behaviours. When human behaviour is related to emotion, these behavioural sounds can be used for emotion detection. In this study, a convolutional neural network was trained to recognise seven different emotional behaviours, based on a newly collected data set. The classification performance was evaluated in terms of class-wise F1 scores, which ranged between 0.54 and 0.90. The study demonstrates the feasibility of using behavioural sounds for emotion detection and gives some first guidelines for implementation. In particular behaviours with regular patterns and impacts on hard surfaces lend themselves well to detection, and the model performs better when the distance between the recording device and the person is short.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page