Epileptic Seizure Detection on EEG Images Using the Decimal Descriptor Pattern

Electroencephalography (EEG) is a widely used tool for the detection of epileptic seizures. However, the visual analysis of long-term EEG recordings is subjective, time-consuming, and prone to errors. To address these issues, various algorithms for epileptic seizure detection have been proposed. In this study, a new automatic seizure detection method is introduced. The feature extraction step is performed using the Decimal Descriptor Pattern (DDP) approach, which is applied here for the first time on 2D images. Subsequently, a Support Vector Machine classifier is used to discriminate between seizure and non-seizure EEG 2D images. The performance of our proposal is evaluated through rhythmicity spectrograms (2D images) from a publicly available EEG database (CHB-MIT Scalp EEG database). The obtained results show high performance in terms of accuracy. Therefore, the proposed method for seizure detection using 2D images can be considered a valuable tool for seizure detection, able to reduce the burden of visual analysis and expedite the seizure diagnosis.