UCTECG-Net: Uncertainty-aware Convolution Transformer ECG Network for Arrhythmia Detection
#UCTECG‑Net #Uncertainty‑aware #Convolution Transformer #ECG classification #Arrhythmia detection #MIT‑BIH #PTB Diagnostic #Deep learning #Spectrogram #Transformer encoder #1D convolution #Model reliability #Safety-critical systems
📌 Key Takeaways
- Deep learning improves ECG classification but lacks clear prediction reliability.
- UCTECG‑Net integrates 1D convolutional layers with Transformer encoders for joint raw‑signal and spectrogram processing.
- The architecture is uncertainty‑aware, providing confidence estimates alongside predictions.
- Evaluation was performed on two benchmark arrhythmia datasets: MIT‑BIH Arrhythmia and PTB Diagnostic.
- Results show superior detection performance compared to existing models.
📖 Full Retelling
The University of Cape Town's researchers have introduced UCTECG-Net, an uncertainty‑aware hybrid model that fuses one‑dimensional convolutions with Transformer encoders to analyze both raw ECG signals and their spectrograms, aiming to improve the reliability of machine‑learning predictions in arrhythmia detection, and they evaluated its performance on the MIT‑BIH Arrhythmia and PTB Diagnostic datasets in 2023.
🏷️ Themes
Artificial Intelligence in Healthcare, Medical Signal Processing, Deep Learning for ECG Analysis, Uncertainty Quantification, Safety‑Critical Decision Making
Entity Intersection Graph
No entity connections available yet for this article.
Original Source
arXiv:2602.16216v1 Announce Type: cross
Abstract: Deep learning has improved automated electrocardiogram (ECG) classification, but limited insight into prediction reliability hinders its use in safety-critical settings. This paper proposes UCTECG-Net, an uncertainty-aware hybrid architecture that combines one-dimensional convolutions and Transformer encoders to process raw ECG signals and their spectrograms jointly. Evaluated on the MIT-BIH Arrhythmia and PTB Diagnostic datasets, UCTECG-Net out
Read full article at source