H-Infinity Filter Enhanced CNN-LSTM for Arrhythmia Detection from Heart Sound Recordings

TL;DR

This work addresses arrhythmia detection from heart sound recordings under challenging noise and data-scarcity conditions. It introduces a CNN-$H_ Infty$-LSTM architecture, where the traditional forget gate is replaced by a trainable $H_ Infty$-inspired mechanism, and couples it with Penalty Weighted Loss and Stochastic Adaptive Probe Thresholding to improve robustness and class-imbalance handling. On the PhysioNet CinC 2016 dataset, the approach achieves a test accuracy of $99.42\%$ and an F1 score of $98.85\%$, outperforming prior audio and vision baselines. The end-to-end design is suitable for scalable deployment on mobile/edge devices, with potential for further improvements via integration with more advanced architectures and adaptive thresholding strategies.

Abstract

Early detection of heart arrhythmia can prevent severe future complications in cardiac patients. While manual diagnosis still remains the clinical standard, it relies heavily on visual interpretation and is inherently subjective. In recent years, deep learning has emerged as a powerful tool to automate arrhythmia detection, offering improved accuracy, consistency, and efficiency. Several variants of convolutional and recurrent neural network architectures have been widely explored to capture spatial and temporal patterns in physiological signals. However, despite these advancements, current models often struggle to generalize well in real-world scenarios, especially when dealing with small or noisy datasets, which are common challenges in biomedical applications. In this paper, a novel CNN-H-Infinity-LSTM architecture is proposed to identify arrhythmic heart signals from heart sound recordings. This architecture introduces trainable parameters inspired by the H-Infinity filter from control theory, enhancing robustness and generalization. Extensive experimentation on the PhysioNet CinC Challenge 2016 dataset, a public benchmark of heart audio recordings, demonstrates that the proposed model achieves stable convergence and outperforms existing benchmarks, with a test accuracy of 99.42% and an F1 score of 98.85%.

Figures (11)

• Figure 1: Waveform representation of the cardiac cycle phases
• Figure 2: Healthy (Top) and Arrhythmic (Bottom) Waveforms
• Figure 3: Effect of wavelet and IIR Filter on the original signal
• Figure 4: Normal signal
• Figure 5: Filtered signal