EfficientECG: Cross-Attention with Feature Fusion for Efficient Electrocardiogram Classification

TL;DR

This work tackles the need for accurate and efficient ECG classification in the presence of multi-lead data and external patient features. It introduces EfficientECG, a lightweight EfficientNet-based backbone adapted with 1D convolutions and MBConv blocks, augmented by R-peak/P-wave feature engineering and LSTM autoencoding. A cross-attention-based multi-feature fusion module integrates age and gender embeddings with ECG representations, achieving superior performance on MIT-BIH, PhysioNet CinC, and HMIC datasets while using substantially fewer parameters than competing approaches. The results demonstrate that multi-feature fusion via cross-attention can meaningfully improve ECG classification, with implications for real-time, resource-constrained diagnostic systems.

Abstract

Electrocardiogram is a useful diagnostic signal that can detect cardiac abnormalities by measuring the electrical activity generated by the heart. Due to its rapid, non-invasive, and richly informative characteristics, ECG has many emerging applications. In this paper, we study novel deep learning technologies to effectively manage and analyse ECG data, with the aim of building a diagnostic model, accurately and quickly, that can substantially reduce the burden on medical workers. Unlike the existing ECG models that exhibit a high misdiagnosis rate, our deep learning approaches can automatically extract the features of ECG data through end-to-end training. Specifically, we first devise EfficientECG, an accurate and lightweight classification model for ECG analysis based on the existing EfficientNet model, which can effectively handle high-frequency long-sequence ECG data with various leading types. On top of that, we next propose a cross-attention-based feature fusion model of EfficientECG for analysing multi-lead ECG data with multiple features (e.g., gender and age). Our evaluations on representative ECG datasets validate the superiority of our model against state-of-the-art works in terms of high precision, multi-feature fusion, and lightweights.

Figures (11)

• Figure 1: The architecture of MBConv.
• Figure 2: The architecture of EfficientECG. The raw ECG records are firstly preprocessed by bio-signal feature engineering, the obtained P-wave and R-peak feature are extracted by LSTM AutoEncoder, and filtered ECG signals are passed to a customized EfficientNet, the three extracted features are combined as the input of classification.
• Figure 3: The distribution of two arrhythmias in various ages and genders.
• Figure 4: The architecture of multi-feature fusion in EfficientECG.
• Figure 5: The architecture of cross-attention module for multi-feature.