| KEYWORD |
Machine Learning and data fusion of physiological signals for assessing a subject's stress level and cognitive load
Thesis in external company
keywords BIOSIGNAL ANALYSIS, DATA FUSION, MACHINE LEARNING, ARTIFICIAL NEURAL NETWORKS
Reference persons DANILO DEMARCHI
External reference persons Irene Buraioli (irene.buraioli@polito.it)
Thesis type INDUSTRIAL
Description Stress and cognitive load are conditions that alter our cognitive and emotional processes, hampering the decision-making process in many situations in our daily lives. Our body is sensitive to such conditions and, in fact, several physiological signals such as brain oxygenation (fNIRS signal), cardiorespiratory signal, EDA signal, body temperature, and eye movements contain within them important information regarding stress and cognitive load. Given the large number of data and the complexity of the problem, to extrapolate this information, it is necessary to study the correlation of these signals with stress using machine learning and artificial intelligence algorithms.
The proposed thesis work will be structured as follows:
1) The candidate will be asked to perform a state-of-the-art analysis of machine learning models applied to the multimodal analysis of physiological signals.
2) Next, the candidate will have to implement a model, identified from the previous analysis, that allows the processing of physiological data correlated with different levels of stress and cognitive load available in a dataset already acquired by our research group.
3) Finally, the candidate will have to validate the functioning of his or her algorithm on data obtained by using ad hoc tests.
This process will allow the candidate to demonstrate the correlation between the aforementioned physiological parameters and the variation of a stress condition and cognitive load by implementing artificial intelligence models on real data.
Required skills Biomedical signal processing, programming in Python, programming in Matlab
Deadline 21/11/2023
PROPONI LA TUA CANDIDATURA