ScaHybNet: a scalogram-based hybrid ensemble network for ECG arrhythmia classification
Sonam Nagar, Karan Verma, Sachin Singh, Kumar Shashvat · 2026
Read ScaHybNet: a scalogram-based hybrid ensemble network for ECG arrhythmia classification by Sonam Nagar, Karan Verma, Sachin Singh, Kumar Shashvat (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13473477.txt.
Description
Sonam Nagar and three colleagues built in 2026 a deep learning ensemble that sorts heartbeats into arrhythmia classes from ECG recordings. Heart disease leads global deaths, and catching abnormal rhythms early can prevent sudden cardiac death. The team used a standard heartbeat dataset of 109,446 samples across five classes: normal, supraventricular ectopic, ventricular ectopic, fusion, and unknown beats, with severe imbalance between classes.
Each signal was first turned into a 224 by 224 colour scalogram image using a wavelet transform. A hybrid network then read it, combining a residual convolutional network, a bidirectional LSTM for time patterns, and a transformer encoder for long-range links. Balanced sampling and class weighting handled the imbalance, and five-fold cross-validation checked robustness. Test accuracy reached 94.73 percent, though precision of 76.51 percent and an F1 score of 77.40 percent show rarer beats still challenge the model.
Catalog record
| CREATOR | Sonam Nagar, Karan Verma, Sachin Singh, Kumar Shashvat |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 159 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |