Hybrid Deep Learning Model for ECG Arrhythmia Detection: Integrating CNN, Quantum Computing and LSTM

View at publisher Summarise with VIRAT
Authors: Sreya Mynampati, Premanand V, D. Girishkanna
Cited by: 0
References: 8

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Electrocardiogram (ECG) signals are very important in the diagnosis of cardiovascular diseases like arrhythmias. Current deep learning models would either concentrate on spatial or temporal feature extraction and thus may fail to grasp the intricate dynamics of ECG data effectively. This work proposes a hybrid deep model that combines the CNNs, QNNs, and LSTM networks for enhanced ECG arrhythmia classification. The model proposed begins by preprocessing PTB-XL 12-lead ECG signals via resampling, noise removal, normalization, and segmentation. CNN layers learn spatial features like QRS morphology, and LSTM layers learn temporal dependencies in the sequence of heartbeats. A quantum layer-parameterized quantum circuits modeled-is introduced between CNN and LSTM phases to facilitate more powerful feature transformation via quantum entanglement and superposition, allowing richer representations for classification. Fully connected layers in the end do multi-class arrhythmia predictions. The hybrid model is compared on the basis of accuracy, precision, recall, F1-score, and AUC across various diagnostic classes like myocardial infarction, conduction disturbances, and ST/T abnormalities. The model proposed in the paper exhibits more classification accuracy and stability than in the case of conventional CNN, LSTM, or Transformer-based models. The point to note here is that by adding QNN, performance improves in edge cases with minimal variations in waveform arrhythmias. This model presents an innovative architectural contribution using quantum feature embedding coupled with classical deep learning to perform biomedical signal classification. It is scalable to low-power hardware implementation, and therefore, it is appropriate for use in wearable health monitoring system.

Fields of Study & Subjects

ECG Monitoring and Analysis Cardiac electrophysiology and arrhythmias Non-Invasive Vital Sign Monitoring

Publication details

Journal / source
International Conference on Emerging Trends and Innovations in ICT (ICEI)
Pages
1–6
Publication date
9 January 2026
Language
EN