Fig 1.
Workflow diagram of the epilepsy seizure detection system architecture.
Fig 2.
Visualization of EEG signals for three types of epilepsy.
Fig 3.
Box plot of the BONN dataset features.
Fig 4.
Box plot of features of epilepsy dataset from NDSC.
Table 1.
Classification tasks for the BONN epilepsy EEG dataset.
Table 2.
Classification tasks for the NDSC epilepsy EEG dataset.
Fig 5.
Performance comparison of different classifiers in the epilepsy detection task.
(a) Three detection tasks on the BONN dataset; (b) Three detection tasks on the NDSC dataset.
Fig 6.
Confusion matrix of models in two datasets detection tasks.
(a)~(c) results of different classifiers on ABCD-E classification task; (d)~(f) results of different classifiers on interictal vs preictal vs ictal.
Fig 7.
SHAP plot analysis of the BONN dataset and NDSC dataset.
(a) BONN dataset (b) NDSC dataset, I stands for interictal, P for preictal, and S for ictal.
Fig 8.
Feature importance analysis of epilepsy dataset from BONN and NDSC.
(a) BONN dataset (b) NDSC dataset, I stands for Interictal, P for Preictal, and S for Ictal.
Table 3.
Results of the cross-dataset epileptic seizure detection.
Table 4.
Comparison of the latest methods on the BONN dataset.