Institutional Repository of Vilnius University
Žmogaus akių mirksnių atpažinimas su neuroniniais tinklais iš žmogaus smegenų elektroencefalografijų /
Abstract
dc:descriptionIn this study we explore which neural network architecture is most suitable for recognition of eye blinks in human electroencephalography data, collected using electroencephalograph provided by UAB “Neurotechnology”. Three distinct neural network architectures, identified through a comprehensive literature review, were compared in two separate experiments. Data collection program which communicates with the EEG device was improved with a new feature and used to collect a data set with 320 five second long EEG recordings from eight different people. The dataset was augmented and preprocessed to mitigate noise and enhance the robustness of the models. Aſter analyzing the results of the experiments a neural network architecture with an accuracy above 75% in both experiments was chosen to be the most suitable. Experiments also showed that there is no need to use EEG data from all eight electrodes when training a neural network, only the ones closest to the eyes are needed. The author also emphasizes that the used electroencephalograph does not have electrodes in all regions of the head and states that it would be valuable to examine the possibilities of the neural network architecture with data collected by other electroencephalographs with more electrodes.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Sipavičius, Dominykas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD210647990&prefLang=en_US