Estimating Similarity between Individual EEG Datasets Using a Convolutional Neural Network

Авторы: Kozyrskiy B.L., Ovchinnikova A.O., Shishkin S.L.
Сборник: Proceedings - 2018 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2018
Год издания: 2019
Место издания: Institute of Electrical and Electronics Engineers Inc
Первая страница: 96
Последняя страница: 101
DOI: 10.1109/SMC.2018.00026
Аннотация: In existing brain-computer interfaces (BCIs) a mental state classifier typically should be trained individually for every user. This means that a user has to perform rather tedious and time-consuming mental tasks before starting to use a BCI. Moreover, amount of train data that can be recorded from a single user is strongly limited, thus limiting quality of classifier training. The range of brain signal variations between the users, however, is not infinite: background and task-related signal patterns can be similar within certain groups of the users. In this work we propose a method for finding users whose data look similar from the classifier’s ’point of view’. It allows for concatenating data from the users demonstrating similarity between their signals, so that larger training sets can be formed, or a classifier trained on one user can be applied in training of a classifier for another user. © 2018 IEEE.
Добавил в систему: Шишкин Сергей Львович

	ИСТИНА	Войти в систему Регистрация
	ИСТИНА ИНХС РАН
	Главная Поиск Статистика О проекте Помощь

ИСТИНА