Control of a Virtual Hand Using a Brain-Computer Interface System: Finding Suitable Features and Classifiers for an Asynchronous BCI System in the Offline State to Control a Virtual Hand in 6-DOF - Tapa blanda

Aziz, Nida

 
9783844304411: Control of a Virtual Hand Using a Brain-Computer Interface System: Finding Suitable Features and Classifiers for an Asynchronous BCI System in the Offline State to Control a Virtual Hand in 6-DOF
Tapa blanda
ISBN 10: 384430441X ISBN 13: 9783844304411
Editorial: LAP LAMBERT Academic Publishing, 2011

Sinopsis

An Asynchronous BCI was designed using non-cue based data collected from several participants in two separate sessions with different protocols. Several different features were extracted and evaluated using the DBI and the best of these, the joint time- frequency (JTF) feature, was chosen for detecting onset and classifying mental tasks. For onset detection, two different methods were employed and compared on the basis of classification performance obtained from confusion matrices, repeatability and ease of application. Different classifiers were tested for onset detection of which Linear Discriminant Analysis (LDA) classifier showed better results. The classified onset using JTF features and LDA classifier was then used to select the active period data in both sessions, which was then sent to the mental task classifier. Mental Task classification was done using Neural Networks, Support Vector Machines and LDA. The performance of the mental task classifier was also evaluated using confusion matrices. A virtual hand was developed graphically in MATLAB and was programmed to move in 6 different directions according to the classified outputs.

"Sinopsis" puede pertenecer a otra edición de este libro.

Reseña del editor

An Asynchronous BCI was designed using non-cue based data collected from several participants in two separate sessions with different protocols. Several different features were extracted and evaluated using the DBI and the best of these, the joint time- frequency (JTF) feature, was chosen for detecting onset and classifying mental tasks. For onset detection, two different methods were employed and compared on the basis of classification performance obtained from confusion matrices, repeatability and ease of application. Different classifiers were tested for onset detection of which Linear Discriminant Analysis (LDA) classifier showed better results. The classified onset using JTF features and LDA classifier was then used to select the active period data in both sessions, which was then sent to the mental task classifier. Mental Task classification was done using Neural Networks, Support Vector Machines and LDA. The performance of the mental task classifier was also evaluated using confusion matrices. A virtual hand was developed graphically in MATLAB and was programmed to move in 6 different directions according to the classified outputs.

Biografía del autor

The author was born in Pakistan. She did her graduate study in Mechatronics Engineering from National University of Science and Technology(NUST), Pakistan and her Postgraduate study in Robotics from UK. Currently she is a Lecturer at NUST.

"Sobre este título" puede pertenecer a otra edición de este libro.

Editorial
LAP LAMBERT Academic Publishing
Año de publicación
2011
Idioma
Inglés
ISBN 10 
384430441X
ISBN 13 
9783844304411
Encuadernación
Tapa blanda
Número de páginas
104
Contacto del fabricante
no disponible
Persona responsable
no disponible