Year: 2025
Type: article
Abstract: Epileptic seizure is a neurological disorder which can create severe consequences when not monitored and given timely care.Various techniques have been proposed for detection of epileptic seizures using electroencephalogram (EEG) signals.But these unimodal approaches have higher false positives.Multimodal approaches combining electrocardiogram (ECG), facial cues etc. with EEG can reduce the false positives.This work proposes a deep learning fusion-based technique for detection of epileptic seizure from multimodal inputs of ECG and EEG.The solution extracts feature from multimodal inputs and applies cross modal learning in spatial and temporal context to increase the accuracy of epileptic seizure detection and reduce false positives.Through experimental analysis with EPILEPSIAE dataset and TUSZ dataset, the proposed solution is found to increase accuracy by at least 1% and reduce false positives by at least 1% compared to recent multimodality solution combining ECG and EEG modality with CNN features and late decision fusion for seizure detection.
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: hybrid
OpenAlex ID: W4407928054