Discrimination of real and imaginary lower body movement: a Deep Learning approach

Manara, Christina; Μαναρά, Χριστίνα

dc.contributor.author	Manara, Christina	en
dc.contributor.author	Μαναρά, Χριστίνα	el
dc.date.accessioned	2024-07-08T08:00:19Z
dc.date.available	2024-07-08T08:00:19Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/59814
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.27510
dc.description	Εθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) "Επιστήμη Δεδομένων και Μηχανική Μάθηση"	el
dc.rights	Default License
dc.subject	Topographical brain maps	en
dc.subject	R3DCNN	en
dc.subject	Optimisers	en
dc.subject	Data augmentation	en
dc.subject	Τοπογραφικοί θερμικοί χάρτες εγκεφάλου	el
dc.subject	Νευρωνικά Δίκτυα	el
dc.subject	Επαναληπτικά Νευρωνικά ∆ίκτυα	el
dc.subject	∆ίκτυα Συνέλιξης	el
dc.subject	Βελτιστοποιητές	el
dc.subject	Deep Learning	en
dc.title	Discrimination of real and imaginary lower body movement: a Deep Learning approach	en
heal.type	masterThesis
heal.classification	Biomedical Engineering	en
heal.classification	Machine Learning	en
heal.language	el
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2024-03-01
heal.abstract	This study introduces a cutting-edge method for analyzing topographical maps derived from electroencephalogram (EEG) data to classify leg movements. Leveraging the spatial information encoded in EEG topographic maps, we propose a hybrid model combining Convolutional Neural Networks (CNNs) with Recurrent Neural Networks (RNNs). This ap- proach is designed to extract and integrate spatial features from the topographic maps and temporal dynamics of EEG signals, respectively. By applying preprocessig techniques (data augmentation, ensemble method for dataset imbalance etc), enhancing the model’s ability to capture the nuanced patterns associated with different leg movements. Differ- ent optimizers, such as Adam, RMSprop & SGD, with different parameters, are performed in order to detect the best model’s performance. Preliminary results show the model’s efficacy in differentiating between specific leg movement tasks, indicating its potential utility in neurorehabilitation and brain-computer interface applications. Our research highlights the significance of advanced signal processing and machine learning tech- niques in interpreting complex brain signals, suggesting avenues for further exploration in optimizing model architecture and improving real-time prediction capabilities.	en
heal.advisorName	Matsopoulos, George	en
heal.committeeMemberName	Tsanakas, Panagiotis	en
heal.committeeMemberName	Georgios, Stamou	en
heal.committeeMemberName	George, Matsopoulos	en
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών. Τομέας Συστημάτων Μετάδοσης Πληροφορίας και Τεχνολογίας Υλικών. Εργαστήριο Βιοϊατρικής Τεχνολογίας	el
heal.academicPublisherID	ntua
heal.numberOfPages	85 σ.	el
heal.fullTextAvailability	false