期刊名称：Signal & Image Processing : An International Journal (SIPIJ)
印刷版ISSN：2229-3922
电子版ISSN：0976-710X
出版年度：2012
卷号：3
期号：1
页码：115
出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：In this paper, we have proposed a new tree based wavelet transform (TBWT) for feature extraction schemefor epileptic seizure detection. Also this paper uses the Directed Acyclic Graph Support Vector Machine(DAGSVM) for the multi-class electroencephalogram (EEG) signals classification. The main aim was todetermine the effective features for this problem. Wavelets have played an important role in biomedicalsignal processing for its ability to capture localized spatial-frequency information of EEG signals. TheTBWT works well for high dimensional, multi-class data streams. Decision making was performed in twostages: feature extraction by computing the approximate and detailed wavelet coefficients andclassification using the classifiers trained on the extracted features. We have compared the TBWT withwavelet based transform by evaluating with the benchmark EEG dataset. Our experimental results showthat the TBWT with DAGSVM gives higher classification accuracy such as 97% than the existing classifier.
关键词：Electroencephalogram (EEG) signals classification; epileptic seizure detection; Tree based Wavelet;Transform; Directed Acyclic Graph Multi-Class Support Vector Machine (DAGSVM)
