DocumentCode :
690565
Title :
Early Identification of Dyslexic Preschoolers Based on Neurophysiological Signals
Author :
Karim, I. ; Qayoom, Abdul ; Wahab, Abdul ; Kamaruddin, Norhaslinda
Author_Institution :
Dept. of Comput. Sci., Int. Islamic Univ. Malaysia, Kuala Lumpur, Malaysia
fYear :
2013
fDate :
23-24 Dec. 2013
Firstpage :
362
Lastpage :
366
Abstract :
Dyslexia is a learning difficulty and in most cases cannot be identified until a child is already in the third grade or later. At this time a dyslexic child have only an one-in-seven chance of ever catching up with his or her peers in reading, writing, speaking or listening. Early identification can pave the way for early intervention and the dyslexic child can be helped at an early stage. Furthermore, the results yielded are the best when the intervention in the form of providing specialized instructions or carried out through some other way yields best results when done at preschoolers. Thus the importance of early identification. The following study is devoted to the EEG based identification of dyslexia for preschool going children. In this analysis feature extraction are carried out using KDE and MLP is used for classification of the features extracted. The results show promising classification accuracy.
Keywords :
electroencephalography; feature extraction; medical signal processing; neurophysiology; paediatrics; signal classification; EEG based identification; KDE; Kernel density estimation; MLP; dyslexic child; dyslexic preschooler identification; feature extraction; learning difficulty; multilayer perceptron; neurophysiological signals; signal classification; Accuracy; Biological neural networks; Computers; Electrodes; Electroencephalography; Feature extraction; Kernel; Dyslexia; EEG; KDE; MLP; Presecreening;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Advanced Computer Science Applications and Technologies (ACSAT), 2013 International Conference on
Conference_Location :
Kuching
Type :
conf
DOI :
10.1109/ACSAT.2013.78
Filename :
6836607
Link To Document :
بازگشت