Enhanced Detection of Inter-Ictal Activity Amidst Artifacts: A Comparative Study Utilizing Deep Learning Frameworks and Ensemble Strategies
Procedia Computer Science · 2026 · 283 · 2577-2586
DOI: 10.1016/j.procs.2026.06.327 ↗Kurz gefasst
Fachartikel in der Zeitschrift Procedia Computer Science, 2026. Verfasst von Arshpreet Kaur und 2 weiteren Autor:innen. Die Arbeit nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Epilepsy is a neurological disorder that has seen significant advancements in seizure detection methodologies with the advent of artificial intelligence. Electroencephalography (EEG) still remains prominent diagnostic tool for epilepsy diagnosis, enabling seizure classification, focus localization, and treatment monitoring. Detecting inter-ictal activity amidst artifacts remains a significant challenge in EEG analysis. Introduction: This study evaluates the performance of a compact Convolutional Neural Network (CNN) against an Ensemble Method that combines CNN-extracted features with a Support Vector Machine (SVM) classifier. Motivated by the promising results of compact CNNs on simpler datasets, the study aimed to accurately distinguish artifacts from inter-ictal and controlled EEG activities. Two cases were analysed: Case 1 compared Controlled vs. (Artifacts vs Inter-Ictal), and Case 2 compared Controlled vs. Artifacts vs. Inter-Ictal. The data were collected from Max Hospital, Saket, Delhi, with all necessary ethical approvals, and model performance was evaluated using K-fold cross-validation and testing sets. The Ensemble Method outperformed the standalone CNN, achieving higher accuracy (95.30% vs. 93.00%) and F1 Score (97.22% vs. 95.80%), demonstrating superior classification of inter-ictal states. These results indicate that ensemble approaches combining CNN features with SVM provide improved accuracy and reliability for EEG-based epilepsy diagnostics.
Zitieren
APA
Kaur, A., Shashvat, K., & Jain, K. (2026). Enhanced Detection of Inter-Ictal Activity Amidst Artifacts: A Comparative Study Utilizing Deep Learning Frameworks and Ensemble Strategies. Procedia Computer Science, 283, 2577-2586. https://doi.org/10.1016/j.procs.2026.06.327
Vancouver
Kaur A, Shashvat K, Jain K. Enhanced Detection of Inter-Ictal Activity Amidst Artifacts: A Comparative Study Utilizing Deep Learning Frameworks and Ensemble Strategies. Procedia Computer Science. 2026;283:2577-2586. doi:10.1016/j.procs.2026.06.327
BibTeX
@article{Kaur2026,
author = {Kaur, Arshpreet and Shashvat, Kumar and Jain, Khushboo},
title = {Enhanced Detection of Inter-Ictal Activity Amidst Artifacts: A Comparative Study Utilizing Deep Learning Frameworks and Ensemble Strategies},
journal = {Procedia Computer Science},
year = {2026},
volume = {283},
pages = {2577--2586},
doi = {10.1016/j.procs.2026.06.327}
}
EUVIMED