Epileptic Seizure Detection Using Convolutional Neural Network: A Multi-Biosignal study

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearch

5 Citations (Scopus)


Epilepsy affects over 70 million people worldwide, making it one of the most common serious neurological disorders in the world. The automated identification of seizures based on EEG signal is one of the most common methods but facing challenges such as the variability of seizures between individual patients and artifact generated during the measurement. In this work, we implement the multi-biosignals scheme for seizure detection by combing EEG, ECG and respiratory. We apply 1D and 2D convolutional neural network (CNN) on multi-biosignal epileptic seizure detection using the in-situ dataset with artifacts. The experimental results show that incorporating multi-biosignals outperforms than using EEG only. We also discovered that Conv2D model could achieve the best AUC of 65%, which is 7% better than the Conv1D model.

Original languageEnglish
Title of host publicationProceedings of the Australasian Computer Science Week Multiconference 2020, ACSW 2020
Editors Abdur Forkan
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Number of pages8
ISBN (Electronic)9781450376976
Publication statusPublished - 4 Feb 2020
EventAustralasian Workshop on Health Informatics and Knowledge Management (HIKM) 2020 - Melbourne, Australia
Duration: 3 Feb 20207 Feb 2020
Conference number: 13th


WorkshopAustralasian Workshop on Health Informatics and Knowledge Management (HIKM) 2020
Abbreviated titleHIKM 2020


  • CNN
  • Epilepsy
  • Multi-Biosignal
  • Multimodal learning
  • Seizure detection

Cite this