This paper presents novel time-frequency (t-f) features based on t-f image descriptors for the automatic detection and classification of epileptic seizure activities in EEG data. Most previous methods were based only on signal-related features derived from the instantaneous frequency and the energies of EEG signals generated from different spectral sub-bands. The proposed features are extracted from the t-f representation of EEG signals which are processed as a textured image using Haralick's texture descriptors. The proposed descriptors are capable to describe visually the epileptic seizure patterns observed in t-f image of EEG signals. The results obtained on real EEG data show that the use of the proposed features improves significantly the performance of the EEG seizure detection and classification by achieving a total classification accuracy up to 99% for 140 EEG segments using one-againt-one SVM classifier.
|Title of host publication||2014 26th International Conference on Microelectronics (ICM)|
|Place of Publication||Piscataway, NJ|
|Publication status||Published - Dec 2014|
|Event||26th International Conference on Microelectronics - Doha|
Duration: 1 Dec 2014 → …
|Conference||26th International Conference on Microelectronics|
|Period||1/12/14 → …|