Dr. Maytham Nabil Miqdad, Department Rapporteur of the Intelligent Medical Systems Department at Al-Mustaqbal University, has published a research paper entitled:
“Enhancing Single-Lead ECG Arrhythmia Classification via Multi-Teacher Decomposed Feature Distillation.”
The paper appears in the reputable journal Computer Methods in Biomechanics and Biomedical Engineering, published by Taylor & Francis – United Kingdom.
This journal is recognized among prestigious scientific outlets, ranked Q2 in Scopus, with a CiteScore of 4.2, and indexed in Clarivate – Web of Science.
🔗 Link to the article on the journal’s website
The research was conducted in collaboration with:
Dr. Fardin Abdali-Mohammadi, Razi University – Kermanshah, Iran
Dr. Majid Sepahvand, Arak University – Markazi Province, Iran
Research Abstract
When a cardiac arrhythmia occurs, all electrocardiogram (ECG) signals provide evidence of it; however, the manifestation is more pronounced in some leads than in others. Building on this medical observation, the proposed Knowledge Distillation (KD) model strengthens weaker leads by transferring information from stronger ones.
The model uses single-lead ECG signals for the student network, while the teacher network is trained on twelve-lead ECG signals. Additionally, Tucker Decomposition is employed to decompose the teacher’s feature maps.
Experimental evaluations showed that the student model achieved an accuracy of 96.48% in arrhythmia classification using the Chapman ECG database.
Al-Mustaqbal University – The First University in Iraq